Pdf and Reported If They Occurred

Peer-Reviewed Journal Tracking and Analyzing Disease Trends Pages 893–1056

EDITOR-IN-CHIEF D. Peter Drotman

Associate Editors EDITORIAL BOARD Paul Arguin, Atlanta, Georgia, USA Timothy Barrett, Atlanta, Georgia, USA Charles Ben Beard, Fort Collins, Colorado, USA Barry J. Beaty, Ft. Collins, Colorado, USA Ermias Belay, Atlanta, Georgia, USA Martin J. Blaser, New York, New York, USA David Bell, Atlanta, Georgia, USA Richard Bradbury, Atlanta, Georgia, USA Christopher Braden, Atlanta, Georgia, USA Sharon Bloom, Atlanta, GA, USA Arturo Casadevall, New York, New York, USA Mary Brandt, Atlanta, Georgia, USA Kenneth C. Castro, Atlanta, Georgia, USA Corrie Brown, Athens, Georgia, USA Louisa Chapman, Atlanta, Georgia, USA Charles Calisher, Fort Collins, Colorado, USA Benjamin J. Cowling, Hong Kong, China Michel Drancourt, Marseille, France Vincent Deubel, Shanghai, China Paul V. Effler, Perth, Australia Isaac Chun-Hai Fung, Statesboro, Georgia, USA Anthony Fiore, Atlanta, Georgia, USA Kathleen Gensheimer, College Park, Maryland, USA David Freedman, Birmingham, Alabama, USA Duane J. Gubler, Singapore Richard L. Guerrant, Charlottesville, Virginia, USA Peter Gerner-Smidt, Atlanta, Georgia, USA Scott Halstead, Arlington, Virginia, USA Stephen Hadler, Atlanta, Georgia, USA Katrina Hedberg, Portland, Oregon, USA Matthew Kuehnert, Atlanta, Georgia, USA David L. Heymann, London, UK Nina Marano, Atlanta, Georgia, USA Keith Klugman, Seattle, Washington, USA Martin I. Meltzer, Atlanta, Georgia, USA Takeshi Kurata, Tokyo, Japan David Morens, Bethesda, Maryland, USA S.K. Lam, Kuala Lumpur, Malaysia J. Glenn Morris, Gainesville, Florida, USA Stuart Levy, Boston, Massachusetts, USA Patrice Nordmann, Fribourg, Switzerland John S. MacKenzie, Perth, Australia John E. McGowan, Jr., Atlanta, Georgia, USA Didier Raoult, Marseille, France Jennifer H. McQuiston, Atlanta, Georgia, USA Pierre Rollin, Atlanta, Georgia, USA Tom Marrie, Halifax, Nova Scotia, Canada Frank Sorvillo, Los Angeles, California, USA Nkuchia M. M’ikanatha, Harrisburg, Pennsylvania, USA David Walker, Galveston, Texas, USA Frederick A. Murphy, Bethesda, Maryland, USA Barbara E. Murray, Houston, Texas, USA Senior Associate Editor, Emeritus Stephen M. Ostroff, Silver Spring, Maryland, USA Brian W.J. Mahy, Bury St. Edmunds, Suffolk, UK Marguerite Pappaioanou, Seattle, Washington, USA Managing Editor Johann D. Pitout, Calgary, Alberta, Canada Byron Breedlove, Atlanta, Georgia, USA Ann Powers, Fort Collins, Colorado, USA Mario Raviglione, Geneva, Switzerland Copy Editors Claudia Chesley, Kristina Clark, Karen Foster, David Relman, Palo Alto, California, USA Thomas Gryczan, Shannon O’Connor, Jude Rutledge, Guenael R. Rodier, Geneva, Switzerland Carol Snarey, P. Lynne Stockton, Deborah Wenger Connie Schmaljohn, Frederick, Maryland, USA Production Thomas Eheman, William Hale, Barbara Segal, Tom Schwan, Hamilton, Montana, USA Reginald Tucker Ira Schwartz, Valhalla, New York, USA Bonnie Smoak, Bethesda, Maryland, USA Editorial Assistants Kristine Phillips, Susan Richardson Rosemary Soave, New York, New York, USA Communications/Social Media Sarah Logan Gregory P. Frederick Sparling, Chapel Hill, North Carolina, USA Robert Swanepoel, Pretoria, South Africa Founding Editor Phillip Tarr, St. Louis, Missouri, USA Joseph E. McDade, Rome, Georgia, USA John Ward, Atlanta, Georgia, USA Emerging Infectious Diseases is published monthly by the Centers for Disease Control and J. Todd Weber, Atlanta, Georgia, USA Prevention, 1600 Clifton Road, Mailstop D61, Atlanta, GA 30329-4027, USA. Telephone Jeffrey Scott Weese, Guelph, Ontario, Canada 404-639-1960, fax 404-639-1954, email [email protected]. Mary E. Wilson, Cambridge, Massachusetts, USA

The conclusions, findings, and opinions expressed by authors contributing Use of trade names is for identification only and does not imply to this journal do not necessarily reflect the official position of the U.S. Depart- endorsement by the Public Health Service or by the U.S. Department of Health ment of Health and Human Services, the Public Health Service, the Centers for and Human Services. Disease Control and Prevention, or the authors’ affiliated institutions. Use of EMERGING INFECTIOUS DISEASES is a registered service mark of trade names is for identification only and does not imply endorsement by any the U.S. Department of Health & Human Services (HHS). of the groups named above. All material published in Emerging Infectious Diseases is in the public ∞ Emerging Infectious Diseases is printed on acid-free paper that meets the domain and may be used and reprinted without special permission; proper requirements of ANSI/NISO 239.48-1992 (Permanence of Paper) citation, however, is required.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 June 2017 Invasive Serotype 35B On the Cover Pneumococci Including an Expanding Serotype Switch Ramón y Cajal Lineage, United States, (1852–1934), Calyx of 2015–2016 Held (3676) (detail), circa 1920. Pencil, India S. Chochua et al. 922 ink, ink washes on Related material available online: cardboard or paper. http://wwwnc.cdc.gov/eid/ Digital image courtesy of article/23/6/17-0071_article the Cajal Institute-CSIC, Madrid, Spain Serologic and Molecular Evidence of Vaccinia Virus Circulation among Small Mammals from About the Cover p. 1052 Different Biomes, Brazil J.B. Miranda et al. 931 Related material available online: http://wwwnc.cdc.gov/eid/ article/23/6/16-1643_article

Relative Risk for Ehrlichiosis and Lyme Disease in an Area Synopsis Genomic Analysis of Salmonella Where Vectors for Both enterica Serovar Typhimurium Are Sympatric, DT160 Associated with a New Jersey, USA 14-Year Outbreak, A. Egizi et al. 939 New Zealand, 1998–2012 S.J. Bloomfield et al. 906 Vector tick infection and encounter Sporadic Creutzfeldt- rates suggest gross underreporting Related material available online: Jakob Disease or misreporting of ehrlichiosis in the in 2 Plasma http://wwwnc.cdc.gov/eid/ article/23/6/16-1934_article northeast United States. Product Recipients, United Kingdom Stockpiling Ventilators for Distribution and Quantitative P. Urwin et al. 893 Influenza Pandemics Estimates of Variant Creutzfeldt-Jakob Disease Two cases of sporadic CJD with H.-C. Huang et al. 914 clotting disorders have been Prions in Tissues of Clinical Related material available online: and Asymptomatic Patients identified, but this may represent a http://wwwnc.cdc.gov/eid/ chance event. article/23/6/16-1417_article J.Y. Douet et al. 946

Research 901 Hospital Outbreaks of Middle East Respiratory Syndrome, Daejeon, South Korea, 2015 J.W. Park et al. 898

Related material available online: http://wwwnc.cdc.gov/eid/ article/23/6/16-0120_article

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Outbreak-Related Disease Burden Associated with Consumption of Unpasteurized Cow’s Milk and Cheese, United States, 2009–2014 S. Costard et al. 957 Related material available online: June 2017 http://wwwnc.cdc.gov/eid/ article/23/6/15-1603_article Dispatches

Sustainability of High-Level Isolation Capabilities among US Ebola Treatment Centers J.J. Herstein et al. 965

Clinical and Molecular Characteristics of Human Rotavirus G8P[8] Outbreak Strain, Japan, 2014 935 K. Kondo et al. 968 Related material available online: High Rates of Neutralizing Crimean-Congo Hemorrhagic http://wwwnc.cdc.gov/eid/ Antibodies to Toscana and Fever in Migrant Worker article/23/6/16-0038_article Sandfly Fever Sicilian Viruses Returning from Oman to in Livestock, Kosovo India, 2016 Seoul Virus Infection In Humans, France, 2014–2016 N. Ayhan et al. 989 P.D. Yadav et al. 1005 J.-M. Reynes et al. 973 Congenital Malformations of Rise in Group W Meningococcal Calves Infected with Shamonda Carriage in University Students, Central Nervous System Virus, Southern Japan United Kingdom Brucellosis Granuloma and White Matter Disease in Y. Hirashima et al. 993 N.J. Oldfield et al. 1009 Immunocompromised Patient Brucella neotomae Infection in Related material available online: M. Alqwaifly et al. 978 http://wwwnc.cdc.gov/eid/ Humans, Costa Rica article/23/6/16-1768_article Severe Neurologic Disorders M. Suárez-Esquivel et al. 997 Penicillin Resistance of in 2 Fetuses with Zika Virus Related material available online: Nonvaccine Type Pneumococcus Infection, Colombia http://wwwnc.cdc.gov/eid/ before and after PCV13 article/23/6/16-2018_article J. Acosta-Reyes et al. 982 Introduction, United States Related material available online: C.P. Andam et al. 1012 http://wwwnc.cdc.gov/eid/ Isolated Case of Marburg Virus

article/23/6/16-1702_article Disease, Kampala, Uganda, 2014 Related material available online: http://wwwnc.cdc.gov/eid/ L. Nyakarahuka et al. 1001 article/23/6/16-1331_article Domestic Pig Unlikely Reservoir for MERS-CoV E. de Wit et al. 985

979

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Febrile Respiratory Illness Associated with Human Adenovirus Type 55 in South Korean Military, 2014–2016 H. Yoo et al. 1016 Related material available online: June 2017 http://wwwnc.cdc.gov/eid/ article/23/6/16-1848_article 990 Commentary

Stockpiling Ventilators for Influenza Pandemics M.I. Meltzer, A. Patel 1021

Letters

Epidemiologic Survey of Japanese Encephalitis Virus Infection, Tibet, China, 2015 H. Zhang et al. 1023

High Frequency of Mayaro Virus Endemic Hantavirus in Field Influenza A(H9N2) Virus, IgM among Febrile Patients, Voles, Northern England Myanmar 2014–2015 Central Brazil A.G. Thomason et al. 1033 T.N. Lin et al. 1041 S.Brunini et al. 1025 Related material available online: Related material available online: Related material available online: http://wwwnc.cdc.gov/eid/ http://wwwnc.cdc.gov/eid/ http://wwwnc.cdc.gov/eid/ article/23/6/16-1607_article article/23/6/16-1902_article article/23/6/16-0929_article Measles Cases during Ebola PCR Detection of Mimivirus Ebola Virus Imported from Outbreak, West Africa, D. Raoult et al. 1043 Guinea to Senegal, 2014 2013–2016 D. Ka et al. 1026 F. Colavita et al. 1035 Autochthonous Case of Eosinophilic Meningitis Caused Tick-Borne Encephalitis Angiostrongylus cantonensis by Angiostrongylus cantonensis, Virus in Ticks and Roe Deer, Meningitis and Myelitis, France, 2016 the Netherlands Texas, USA Y. Nguyen et al. 1045 S. Jahfari et al. 1028 R. Al Hammoud et al. 1037 Zika Virus–Associated Cognitive Outbreaks of Tilapia Lake Virus Enterocytozoon bieneusi Impairment in Adolescent, 2016 Infection, Thailand, 2015–2016 Microsporidiosis in Stem Cell J. Zucker et al. 1047 W. Surachetpong et al. 1031 Transplant Recipients Treated Related material available online: with Fumagillin Related material available online: http://wwwnc.cdc.gov/eid/ http://wwwnc.cdc.gov/eid/ I. Bukreyeva et al. 1039 article/23/6/16-2029_article article/23/6/16-1278_article Highly Pathogenic Avian Influenza Virus (H5N8) Clade 983 2.3.4.4 Infection in Migratory Birds, Egypt A.A. Selim et al. 1048 Related material available online: http://wwwnc.cdc.gov/eid/ article/23/6/16-2056_article

About the Cover “Unexplored Continents and Great Stretches of Unknown Territory” B. Breedlove 1052

Etymologia

Creutzfeldt-Jakob Disease R. Henry, F.A. Murphy 956

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 These spotlights highlight the latest articles and information on emerging infectious disease topics in our global community. Antimicrobial Resistance Food Safety Ebola HIV/AIDS Lyme Disease Influenza MERS Pneumonia Rabies Ticks Tuberculosis Zika

httnoveeothtto SYNOPSIS Sporadic Creutzfeldt-Jakob Disease in 2 Plasma Product Recipients, United Kingdom Patrick Urwin, Kumar Thanigaikumar, James W. Ironside, Anna Molesworth, Richard S. Knight, Patricia E. Hewitt, Charlotte Llewelyn, Jan Mackenzie, Robert G. Will

In support of improving patient care, this activity has been planned and implemented by Medscape, LLC and Emerging Infectious Diseases. Medscape, LLC is jointly accredited by the Accreditation Council for Continuing Medical Education (ACCME), the Accreditation Council for Pharmacy Education (ACPE), and the American Nurses Credentialing Center (ANCC), to provide continuing education for the healthcare team. Medscape, LLC designates this Journal-based CME activity for a maximum of 1.00 AMA PRA Category 1 Credit(s)™. Physicians should claim only the credit commensurate with the extent of their participation in the activity. All other clinicians completing this activity will be issued a certificate of participation. To participate in this journal CME activity: (1) review the learning objectives and author disclosures; (2) study the education content; (3) take the post- test with a 75% minimum passing score and complete the evaluation at http://www.medscape.org/journal/eid; and (4) view/print certificate. For CME questions, see page 1,055. Release date: May 12, 2017; Expiration date: May 12, 2018

Learning Objectives

Upon completion of this activity, participants will be able to: • Recognize the clinical features of 2 cases of sporadic Creutzfeldt-Jakob disease (sCJD) reported in patients with clotting disorders treated with fractionated plasma products. • Identify the laboratory and pathology findings of 2 cases of sCJD reported in patients with clotting disorders treated with fractionated plasma product. • Determine the clinical implications of 2 cases of sCJD reported in patients with clotting disorders treated with fractionated plasma products. CME Editor Jude Rutledge, BA, Technical Writer/Editor, Emerging Infectious Diseases. Disclosure: Jude Rutledge has disclosed no relevant financial relationships. CME Author Laurie Barclay, MD, freelance writer and reviewer, Medscape, LLC. Disclosure: Laurie Barclay, MD, has disclosed the following relevant financial relationships: owns stock, stock options, or bonds from Alnylam; Biogen; Pfizer. Authors Disclosures: Patrick Urwin, MBBS, MA (CANTAB); Kumar Thanigaikumar, MBBS, MRCP, FRCPATH; James W. Ironside, MD; Anna Molesworth, PhD; Patricia E. Hewitt, MD, FRCPATH; Charlotte A. Llewelyn, PhD; and Jan Mackenzie, PG Cert Epidemiology, have disclosed no relevant financial relationships. Richard S. Knight, BMBCh, FRCP (E), has disclosed the following relevant financial relationships: served as a speaker or a member of a speakers bureau for Pfizer Inc. Robert G. Will, MD, has disclosed the following relevant financial relationships: served as an advisor or consultant for LFB (Paris); Ferring Pharmaceuticals.

Author affiliations: University of Edinburgh Western General Sporadic Creutzfeldt-Jakob disease (sCJD) has not been Hospital, Edinburgh, Scotland, UK (P. Urwin, J.W. Ironside, previously reported in patients with clotting disorders treat- A. Molesworth, R.S. Knight, J. Mackenzie, R.G. Will); University ed with fractionated plasma products. We report 2 cases of Hospital Lewisham, London, UK (K. Thanigaikumar); National sCJD identified in the United Kingdom in patients with a his- Health Service Blood and Transplant, London (P.E. Hewitt); tory of extended treatment for clotting disorders; 1 patient National Health Service Blood and Transplant/Public Health had hemophilia B and the other von Willebrand disease. England Epidemiology Unit, Cambridge, UK (C. Llewelyn) Both patients had been informed previously that they were at increased risk for variant CJD because of past treatment DOI: https://dx.doi.org/10.3201/eid2306.161884 with fractionated plasma products sourced in the United

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 893

Page 1 of 1 SYNOPSIS

Kingdom. However, both cases had clinical and investiga- typical of sCJD; the neuropathologic diagnosis in both cas- tive features suggestive of sCJD. This diagnosis was con- es was sCJD (subtype MM1). firmed in both cases on neuropathologic and biochemical analysis of the brain. A causal link between the treatment The Investigation with plasma products and the development of sCJD has not The UK National CJD Research and Surveillance Unit has been established, and the occurrence of these cases may been carrying out systematic epidemiologic study of CJD simply reflect a chance event in the context of systematic surveillance for CJD in large populations. since 1990. The methodology of this study has been published previously (7). In brief, patients with suspected CJD are referred by clinicians and visited by a research registrar, uman prion diseases are a group of rare and fatal who obtains details of the clinical history and investiga- Hneurodegenerative diseases that include idiopathic tions, information on a range of possible risk factors, and (sporadic), genetic (inherited), and acquired (infectious) past medical history. The Transfusion Medicine Epidemi- disorders (1). All are associated with the accumulation of ology Review study investigates potential links between an abnormal isoform of the prion protein (PrPSc) in the cen- donors and recipients of labile blood components and, in tral nervous system (1). The most common human prion cases of sCJD, investigates patients who have a history of disease is the sporadic form of Creutzfeldt-Jakob disease blood donation or having received a blood transfusion. (sCJD), which occurs worldwide with a relatively uniform Coordinated surveillance of CJD has been undertaken incidence of 1–2 cases per million population per year, in the European Union since 1993 (8). National surveil- a peak incidence in the 7th decade of life, and a median lance programs for CJD also are in place in several oth- duration of illness of 4 months. The relatively consistent er countries, including Australia, Canada, Japan, and the mortality rates associated with sCJD, the overall random United States. spatial and temporal distribution of cases, and the absence of any confirmed environmental risk factor have led to Case 1 the hypothesis that sCJD occurs because of the spontane- In 2014, a 64-year-old woman suffered a rapidly progres- ous generation of PrPSc in the brain (1). In contrast, variant sive dementia with deterioration in driving skills and bal- Creutzfeldt-Jakob disease (vCJD) is an acquired disorder ance disturbance, then limb coordination deficits with that is most likely caused by the consumption of meat or handwriting impairment. In the second month, her gait de- meat products contaminated with the bovine spongiform teriorated, becoming shuffling and unsteady, she struggled encephalopathy agent. The median age at death in vCJD is to dress herself, and she had onset of daytime hypersomno- 30 years, with a median duration of illness of 14 months. lence. She became distractible, had visual misperceptions, Most cases of vCJD have occurred in the United Kingdom, emotional lability, and spatial memory problems. She was which has had the largest epizootic of bovine spongiform hospitalized at the beginning of the third month of her ill- encephalopathy in the world. Of the 178 UK vCJD cases, ness and had onset of cortical blindness, myoclonus, and 3 have been identified as cases of secondary transmission akinetic mutism. She experienced rapid decline and died caused by the transfusion of nonleukodepleted red blood after a total illness duration of 3 months. cell components from vCJD-infected blood donors. An electroencephalogram performed during the final Lookback studies have shown no evidence of trans- stages of illness showed background slowing and runs of mission through blood transfusion in sCJD (2,3), despite periodic complexes, and a magnetic resonance imaging the identification of ScPrP in some peripheral tissues (4) (MRI) brain scan showed high signal in the caudate heads and experimental evidence, which demonstrated infectiv- with posterior cortical ribboning. A cerebral spinal fluid ity in blood (5) by using intracerebral inoculation of highly (CSF) 14–3–3 assay and real-time quaking-induced con- sensitive transgenic mice. The absence of clinical cases version test for PrPSc both were positive. Prion protein gene causally linked to past treatment with fractionated plasma (PRNP) sequencing showed no mutations with methionine products has been used as evidence of the safety of these homozygosity at codon 129. products in relation to sCJD (6). These products are gen- Postmortem examination of the brain showed wide- erally manufactured from the pooled plasma from several spread spongiform encephalopathy of predominantly mi- thousand donors; production using UK plasma was discon- crovacuolar type. Immunocytochemistry for prion protein tinued in 1999. gave a widespread positive reaction in a granular/synap- We describe 2 cases of sCJD in patients who had tic pattern (Figure). No plaques or plaque-like structures previously received treatment with UK plasma–sourced were identified. Results of immunocytochemistry for dis- plasma products; both patients had been informed that they ease-associated prion protein were negative in peripheral were at increased risk for vCJD because of that treatment. nerve, liver, lymph node, appendix, and spleen. Western The clinical features and investigations in these cases were blot analysis of frontal cortex and cerebellum confirmed

894 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 sCJD in 2 Plasma Product Recipients, UK

Figure. Results of neuropathologic examinations of the brains of the 2 patients with sporadic Creutzfeldt-Jakob disease, United Kingdom, 2014. A) Microvacuolar spongiform change in the frontal cortex (case 1). Hematoxylin and eosin stain; original magnification ×400. B) Fine granular/ synaptic accumulation of abnormal prion protein in the cerebral cortex (case 1). 12F10 antiprion protein antibody; original magnification ×400. C) Microvacuolar spongiform change with neuronal loss and gliosis in the frontal cortex (case 2). Hematoxylin and eosin stain; original magnification ×400. D) Focally intense granular/ synaptic accumulation of abnormal prion protein in the cerebral cortex (case 2). 12F10 antiprion protein antibody; original magnification ×400. the presence of protease-resistant prion protein with a type 12 years before clinical onset in the recipient, died in 2013 1A isoform. at 76 years of age, and the diagnoses on the death certificate The patient had been diagnosed with von Willebrand were vascular dementia and bladder cancer. Identification of disease in childhood. Her early therapies include numer- donors for transfusions before 2001 has not been possible. ous transfusions of red blood cells and platelets; in more recent years, she received plasma-derived and recombinant Case 2 factor VIII and additional blood component transfusions at In 2014, a 64-year-old woman reported day/night reversal times of hemorrhage. Factor VIII was administered on 4 of sleep patterns and, 3 months later, excessive tearfulness, occasions in the 1990s and during 2000–2004 and von Wil- for which she was started on antidepressants. She then had lebrand factor/factor VIII (Haemate-P) during 2001–2013. onset of writing problems, followed during the next few Because of her history of exposure to UK-sourced plasma days by increasing language problems that led to expres- products, for public health purposes she had been informed sive dysphasia. She deteriorated rapidly thereafter, requir- that she was at risk for vCJD, although she was not known ing assistance with her activities of daily living and having to have been exposed to factor VIII derived from a batch coordination and memory problems, jerking movements including a vCJD donation. She had no history of potential suggestive of myoclonus, and itching in both arms. She iatrogenic exposure to CJD and no family history of CJD. was admitted to the hospital and experienced a probable Donors for all blood or platelet transfusions since 2001 focal seizure with secondary generalization. She had onset have been identified. Of the 107 donors, 106 are still alive, of a homonymous hemianopia and limb rigidity and then with a median age of 55 years (range 27–80 years). (Table 1). became bedbound and mute, dying 7 months after the onset One donor of leukodepleted platelets, which were transfused of symptoms.

Table 1. Selected characteristics of blood donors to the patient with sporadic Creutzfeldt-Jakob disease described in case 1, United Kingdom, 2014* Interval from transfusion to onset, y Component No. donors No. donors alive No. donors dead 3 RBC LD 4 4 0 6 RBC LD 6 6 0 7 RBC LD 19 19 0 9 RBC LD 3 3 0 10 RBC LD 4 4 0 12 Whole blood LD; RBC LD; platelets LD 2; 27; 42 2; 27; 41 0; 0; 1 *LD, leukodepleted; RBC, red blood cells. Median age of donors, 56 years (range 27–80 years).

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 895

SYNOPSIS

An electroencephalogram performed during the final of these cases may simply reflect a chance event in the con- stages of illness showed widespread slowing, more evident text of systematic surveillance of CJD in large populations. on the left. An MRI brain scan showed left-sided caudate Both patients had been informed that they were at in- head and anterior putaminal high signal. Diffusion weight- creased risk for vCJD, and considering the evidence for the ed imaging showed areas of cortical high signal. Results of type of CJD in the 2 cases is important. Both patients had a CSF 14–3–3 assay and real-time quaking-induced con- a clinical phenotype suggestive of sCJD, including a short version tests were positive. Consent for full sequencing of duration of illness, typical early symptoms, a suggestive the PRNP was not obtained; methionine homozygosity at MRI scan, and, in 1 patient, a typical EEG. Notably, both codon 129 was identified. patients had a positive real-time quaking-induced conver- Postmortem neuropathologic examination of the brain sion test result for PrPSc in CSF; previously this test had not showed a widespread spongiform encephalopathy with mi- been positive in any case of vCJD evaluated in our labora- crovacuolar spongiform change, neuronal loss, and gliosis. tory (Table 2) (9). However, neuropathological examina- Immunostaining for prion protein showed widespread position was critical; it showed appearances typical of sCJD in tivity with a granular/synaptic pattern (Figure). No amyloid both patients and no evidence of peripheral pathogenesis plaques were identified. Western blot analysis confirmed on immunostaining of lymphoreticular tissues, a feature the presence of protease resistant prion protein with a type that is observed in all tested specimens of vCJD patients 1A isoform. There was no evidence of abnormal prion pro- to date (10). Furthermore, both patients had a type 1A iso- tein accumulation in spleen and appendix either on immu- form PrPSc on Western blot consistent with a diagnosis of nocytochemistry or high sensitivity Western blot analysis. sCJD subtype MM1 (11). Neither patient had a history of The patient was known to have hemophilia B since potential iatrogenic exposure or a family history of CJD, 1964 and had received plasma-derived and recombinant and for the case for which sequencing of the PRNP was factor IX during 1984–2012. For public health purposes, performed, no mutations were detected. In both cases, an she had been informed that she was at risk for vCJD and in MM genotype occurred at codon 129 of PRNP, which does 1991 had received factor IX derived from a pool containing not distinguish between sCJD and vCJD. Laboratory trans- plasma from a donor who subsequently had vCJD. She had mission studies to provide evidence of agent strain in the no history of potential iatrogenic exposure to CJD and no cases have not been possible. family history of CJD. One patient had received multiple transfusions of blood In 1985, the patient received 6 units of fresh frozen components over an extended period, and the other had re- plasma (FFP). Tracing of donors has not been possible. ceived 6 units of FFP 19 years before clinical onset, raising the possibility that these cases could have resulted from sec- Discussion ondary transmission through blood components. In the case This report describes 2 cases of sCJD in patients with a his- of the patient with von Willebrand disease, 107 donors have tory of treatment with UK-sourced plasma products, 1 with been traced, and none appear in the register of cases of CJD a history of hemophilia B and 1 with von Willebrand’s dis- kept at the National CJD Research and Surveillance Unit. ease. To our knowledge, no previous case of sCJD in a per- However, it has not been possible to obtain information son with a history of extended exposure to plasma products on blood transfusions for this patient before 2001 nor on has been reported. It is clearly of concern that there have the FFP transfusions for the patient with hemophilia B. been 2 such cases in a relatively short period in the UK, Lookback studies in the United States and United Kingdom where many plasma product recipients have been informed have provided no evidence of transfusion-transmission of that they are at increased risk for vCJD. However, a causal sCJD (2,3), and although 1 study suggested an increase in link between the treatment with plasma products and the risk after a lag period of 10 years (12), this finding was not onset of sCJD has not been established, and the occurrence confirmed in another study (13). The balance of evidence

Table 2. Selected characteristics and clinical features of the 2 patients with sCJD described in cases 1 and 2, United Kingdom, 2014* Characteristic/clinical feature Case 1 Case 2 Patient age, y/sex 64/F 64/F Symptoms/signs Ataxia, cognitive impairment, visual Somnolence/depression, dysphasia, impairment, myoclonus cognitive impairment, myoclonus/ataxia Magnetic resonance imaging + + Electroencephalogram + Slow activity Cerebrospinal fluid 14–3–3 assay + + RT-QuIC + + Genotype MM MM Diagnosis Definite sCJD Definite sCJD Duration 3 mo 7 mo *RT-QuIC, real-time quaking-induced conversion; sCJD, sporadic Creutzfeldt-Jakob disease.

896 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

sCJD in 2 Plasma Product Recipients, UK indicates that, if sCJD is transmitted by blood transfusion, of Creutzfeldt-Jakob disease in a US surveillance study. it must be a rare event, if it happens at all, and transfusion Transfusion. 2009;49:977–84. http://dx.doi.org/10.1111/ j.1537-2995.2008.02056.x transmission is probably not the explanation for the 2 cases 3. Urwin PJ, Mackenzie JM, Llewelyn CA, Will RG, Hewitt PE. we describe. Creutzfeldt-Jakob disease and blood transfusion: updated results Systematic surveillance for CJD, including a coordi- of the UK Transfusion Medicine Epidemiology Review Study. Vox nated study in Europe (14), has been carried out in many Sang. 2016;110:310–6. http://dx.doi.org/10.1111/vox.12371 4. Glatzel M, Abela E, Maissen M, Aguzzi A. Extraneural pathologic countries over the past 25 years and is continuing. Many of prion protein in sporadic Creutzfeldt-Jakob disease. N Engl J Med. these studies obtain information on potential risk factors, 2003;349:1812–20. http://dx.doi.org/10.1056/NEJMoa030351 including details of past medical history. To date, no case 5. Douet JY, Zafar S, Perret-Liaudet A, Lacroux C, Lugan S, of sCJD has been reported in a person who has received Aron N, et al. Detection of infectivity in blood of persons with variant and sporadic Creutzfeldt-Jakob disease. treatment for a clotting disorder. In fact, the absence of such Emerg Infect Dis. 2014;20:114–7. http://dx.doi.org/10.3201/ a case has been used to argue against the possibility that eid2001.130353 plasma-derived products pose a risk for sCJD transmission 6. European Medicines Agency. CHMP position statement on (6). CJD surveillance centers are aware of the relevance of Creutzfeldt-Jakob disease and plasma-derived and urine-derived medicinal products. London, 23 June 2011 [cited 2015 May 1]. this issue, and sCJD patients with a history of treatment http://www.ema.europa.eu/docs/en_GB/document_library/ with plasma products probably would have been identified Position_statement/2011/06/WC500108071.pdf and reported if they occurred. Although it is surprising that 7. Cousens SN, Zeidler M, Esmonde TF, De Silva R, Wilesmith JW, 2 cases of sCJD have been identified among a population of Smith PG, et al. Sporadic Creutzfeldt-Jakob disease in the United Kingdom: analysis of epidemiological surveillance data for 4,000–5,000 patients in the UK who have been treated for 1970–96. BMJ. 1997;315:389–95. http://dx.doi.org/10.1136/ clotting disorders with fractionated plasma products, the bmj.315.7105.389 total population under surveillance for CJD in Europe and 8. Wientjens DPWM, Will RG, Hofman A. Creutzfeldt-Jakob disease: internationally exceeds 500 million. Assuming an annual a collaborative study in Europe. J Neurol Neurosurg Psychiatry. 1994;57:1285–99. incidence rate of sCJD of 1.5–2.0 per million population 9. McGuire LI, Peden AH, Orrú CD, Wilham JM, Appleford NE, (15), the occurrence of 2 cases of sCJD in this total popula- Mallinson G, et al. Real time quaking-induced conversion analysis tion may not imply a causal link between the treatment and of cerebrospinal fluid in sporadic Creutzfeldt-Jakob disease. Ann the occurrence of the disease. The 2 cases were identified Neurol. 2012;72:278–85. http://dx.doi.org/10.1002/ana.23589 10. Head MW, Ritchie D, Smith N, McLoughlin V, Nailon W, Samad S, over a period of months, and no further cases have been et al. Peripheral tissue involvement in sporadic, iatrogenic, and found since 2014; however, continuing to search for such variant Creutzfeldt-Jakob disease: an immunohistochemical, cases through CJD surveillance programs is essential. quantitative, and biochemical study. Am J Pathol. 2004;164:143– 53. http://dx.doi.org/10.1016/S0002-9440(10)63105-7 11. Head MW, Bunn TJR, Bishop MT, McLoughlin V, Lowrie S, Acknowledgment McKimmie CS, et al. Prion protein heterogeneity in sporadic but We thank Mark W. Head for the Western blot data. not variant Creutzfeldt-Jakob disease: UK cases 1991–2002. Ann Neurol. 2004;55:851–9. http://dx.doi.org/10.1002/ana.20127 The National CJD Research and Surveillance Unit is supported 12. Puopolo M, Ladogana A, Vetrugno V, Pocchiari M. Transmission of by the Policy Research Program of the Department of Health sporadic Creutzfeldt-Jakob disease by blood transfusion: risk factor and the Government of Scotland (grant no. PR-ST-0614-00008). or possible biases. Transfusion. 2011;51:1556–66. http://dx.doi.org/ 10.1111/j.1537-2995.2010.03004.x This report is independent research in part funded by the 13. Molesworth AM, Mackenzie J, Everington D, Knight RSG, Department of Health Policy Research Programme and the Will RG. Sporadic Creutzfeldt-Jakob disease and risk of blood Government of Scotland. The views expressed in this publication transfusion in the United Kingdom. Transfusion 2011;51:1872–3. are those of the authors and not necessarily those of the https://doi.org/10.1111/j.1537-2995.2011.03198.x 14. Ladogana A, Puopolo M, Croes EA, Budka H, Jarius C, Collins S, Department of Health or the Government of Scotland. et al. Mortality from Creutzfeldt-Jakob disease and related disorders Dr. Urwin worked as a research registrar at the National CJD in Europe, Australia, and Canada. Neurology. 2005;64:1586–91. http://dx.doi.org/10.1212/01.WNL.0000160117.56690.B2 Research and Surveillance Unit and is currently training in 15. Minikel EV, Vallabh SM, Lek M, Estrada K, Samocha KE, neurology. His primary research interests include human Sathirapongsasuti JF, et al.; Exome Aggregation Consortium prion diseases. (ExAC). Quantifying prion disease penetrance using large population control cohorts. Sci Transl Med. 2016;8:322ra9. http://dx.doi.org/10.1126/scitranslmed.aad5169 References 1. Prusiner SB. Molecular biology of prion diseases. Science. Address for correspondence: R.G. Will, National CJD Research and 1991;252:1515–22. http://dx.doi.org/10.1126/science.1675487 Surveillance Unit, Western General Hospital, Edinburgh, Scotland EH4 2. Dorsey K, Zou S, Schonberger LB, Sullivan M, Kessler D, Notari E IV, et al. Lack of evidence of transfusion transmission 2XU, UK; email: [email protected]

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 897 RESEARCH Hospital Outbreaks of Middle East Respiratory Syndrome, Daejeon, South Korea, 2015 Jung Wan Park, Keon Joo Lee, Kang Hyoung Lee, Sang Hyup Lee, Jung Rae Cho, Jin Won Mo, Soo Young Choi, Geun Yong Kwon, Ji-Yeon Shin, Jee Young Hong, Jin Kim, Mi-Yeon Yeon, Jong Seok Oh, Hae-Sung Nam

From May through July 2015, a total of 26 cases of Middle The first case of MERS in South Korea was reported East Respiratory Syndrome were reported from 2 hospitals on May 20, 2015. The patient had flown among several in Daejeon, South Korea, including 1 index case and 25 new countries in the Middle East (Bahrein, the United Arab cases. We examined the epidemiologic features of these Emirates, Saudi Arabia, and Qatar) and became the source cases and found an estimated median incubation period of of consecutive hospital-to-hospital transmissions after 6.1 days (8.8 days in hospital A and 4.6 days in hospital B). his return to South Korea, which led to 186 laboratory- The overall attack rate was 3.7% (4.7% in hospital A and 3.0% in hospital B), and the attack rates among inpatients confirmed cases (5) and 38 deaths. Hospital-to-hospital and caregivers in the same ward were 12.3% and 22.5%, re- transmission involved 17 hospitals and originated from 1 spectively. The overall case-fatality rate was 44.0% (28.6% hospital (hospital P) (5,6). This transmission was attribut- in hospital A and 63.6% in hospital B). The use of cohort able to “hospital shopping” by some MERS patients (4,5) quarantine may have played a role in preventing community and was particularly evident in Daejeon, which is the fifth spread, but additional transmission occurred among mem- largest city in South Korea. The index case-patient for bers of the hospital cohort quarantined together. Caregivers nosocomial transmission in Daejeon had initially trav- may have contributed in part to the transmission. eled from his home city of Daejeon to Pyeongtaek, South Korea, seeking healthcare at hospital P, after which he few respiratory viruses constitute emerging threats to returned to Daejeon. Subsequently, 2 hospitals in Daejeon Aglobal health security (1); among them are Middle East experienced MERS cases attributable to this patient. This respiratory syndrome (MERS) coronavirus (MERS-CoV), index case-patient in Daejeon was consecutively hospital- which has caused outbreaks in Saudi Arabia (2,3). The ma- ized at hospital A in Daejeon during May 22–28, 2015, jor MERS outbreaks that occurred during 2012–2015 have and at hospital B during May 28–30, 2015. Thereafter, an been in or near the Arabian Peninsula. However, informa- additional 25 MERS cases (14 in hospital A, 11 in hospi- tion on the epidemiologic features of MERS is insufficient, tal B) were reported. especially for different environmental and cultural settings. After the South Korea government recognized the The 2015 MERS outbreak in South Korea could provide outbreak of MERS in Daejeon, cohort quarantine (isola- more information about the epidemiology of MERS because tion of persons who had been in contact with patients with it was the largest outbreak outside the Middle East (4). confirmed cases in the hospital ward) was applied. This quarantine seems to have played a useful role in preventing Author affiliations: Korea Centers for Disease Control and the spread of MERS-CoV to the local community. We de- Prevention, Cheongju, South Korea (J.W. Park, K.J. Lee); Sokcho scribe the MERS case-patients, the epidemiologic features Medical Center, Sokcho, South Korea (K.H. Lee); Ulsan of the disease, and the quarantine policy used to prevent Metropolitan City Hall, Ulsan, South Korea (S.H. Lee); Daejeon additional transmission. Provincial Government, Daejeon, South Korea (J.R. Cho); Chuncheongbukdo Provincial Government, Cheongju (J.W. Mo); Methods DaeChung Hospital, Daejeon (S.Y. Choi); Ministry of Health and Welfare, Sejong, South Korea (G.Y. Kwon); Eulji University School Setting of Medicine, Daejeon (J.-Y. Shin); Konyang University College of Hospital A is a 300-bed general hospital in Daejeon. The Medicine, Daejeon (J.Y. Hong); Chungnam National University outbreak occurred in ward 51 on the fifth floor, where 13 School of Medicine, Daejeon (J. Kim, M.-Y. Yeon, H.-S. Nam); rooms (5 with 7 beds, 6 with 4 beds, 1 with 2 beds, and 1 Gwanggyo Ean Pediatric Clinic, Suwon, South Korea (J.S. Oh) with 1 bed) are located. Hospital B is an 800-bed university DOI: http://dx.doi.org/10.3201/eid2306.160120 hospital in Daejeon. The main outbreak occurred in ward

898 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Hospital Outbreaks of MERS, South Korea, 2015

101 on the ninth floor, where 16 rooms (7 with 6 beds, 1 viral loads. For each patient with a confirmed case, the Ko- with 4 beds, 1 with 2 beds, and 7 with 1 bed) are located. rea Centers for Disease Control and Prevention assigned a case number according to the order of confirmation during Data Collection and Exposure Assessment the 2015 MERS outbreak in South Korea. For example, the Epidemiologic investigators of the Korea Centers for case number of the index case-patient in Daejeon was 16. Disease Control and Prevention started their outbreak investigation with face-to-face interviews of the index case- Data and Statistical Analyses patient in Daejeon and the 25 additional case-patients The cases were described in case-series form. Attack rates with confirmed MERS-CoV infection. We collected data were calculated as the number of cases per number of ex- on the demographic characteristics and the clinical, con- posed persons (defined as persons who had experienced tact, and MERS-CoV exposure histories and thoroughly face-to-face contact with a symptomatic MERS case-pa- reviewed the medical records of the case-patients to iden- tient in either hospital or as persons who had been in the tify symptoms, underlying concurrent medical conditions, same hospital ward as the symptomatic case-patients). laboratory findings, and clinical courses of illness. Clini- Such persons were identified from the outbreak investi- cal outcome was classified as recovery or death, and the gation reports and the lists of those undergoing cohort or ambulation status of the inpatients at the time of admis- home quarantine. To assess differences in attack rates and sion was clarified. case-fatality rates according to independent variables, we We collected the names of inpatients, their room num- performed χ2 and Fisher exact tests by using SAS software bers, medical staff, and caregivers (family members or version 9.3 (SAS Institute, Inc., Cary, NC, USA). Compari- professionals hired by the family or hospital) exposed to sons were considered significant at p<0.05 and marginally MERS-CoV in each hospital. The duration and route of significant at p<0.1 (both p values were 2-tailed). exposure were further determined by reviewing record- We defined the incubation period as the time from ings from closed-circuit televisions placed in the hospitals. exposure to onset of MERS-associated symptoms, includ- Moreover, we used the floor plan of each hospital to es- ing nonspecific signs and symptoms such as fever, chills, timate the spatial distributions and transmission routes of cough, sore throat, sputum production, dyspnea, myal- the virus within the hospitals. These estimates enabled us gia, headache, nausea, vomiting, diarrhea, and abdominal to identify a possible location of exposure and a transmis- discomfort. If the exposure period was >2 days, a single sion route for each confirmed case. When a patient with interval-censored estimate of the incubation period was a confirmed case had experienced several possible expo- computed by using the earliest and latest dates of exposures, we determined the most probable exposure by author sure and the date of symptom onset for each case-patient consensus. Persons who had had face-to-face contact with (coarseDataTools package in R statistical software version patients with confirmed cases were considered the clos- 3.2.2) (7). To construct cumulative fraction curves of all est contacts. When the data were ambiguous, the follow- cases by incubation period, we calculated the log-normal ing were reviewed independently by the Korea Centers for density function by fitting the interval-censored data on in- Disease Control and Prevention and the Daejeon In-Depth cubation periods. To do this, we used the maximum-likeli- team: all potential exposures by symptom onset; disease hood method and calculated the medians and 5th and 95th duration; physical distance from a patient with a confirmed percentiles of the incubation periods. case; and infector factors including ambulation status, symptoms (including a productive cough), and sharing of Results caregivers. When a patient with a confirmed case had been subjected to several potential exposures, the most probable Description of the Daejeon Index Case-Patient exposure was determined by consensus of the 2 teams. An The Daejeon index case-patient (case-patient 16), a 41-year- expert member of the Korean Society of Epidemiology re- old man, lived in Daejeon and was a former smoker (10 to viewed all decisions. The process was repeated until a final 20 pack-years). He had undergone colon surgery in August consensus was obtained. 2014 at hospital P. The index case-patient of the MERS outbreak in South Korea (case-patient 1) was in hospital P Laboratory Diagnoses during May 15–17, 2015. The Daejeon index case-patient Sputum samples from the persons suspected of having was admitted to hospital P at the same time (May 15–18, MERS were collected in sterile cups and sent to qualified 2015) for a follow-up colonoscopy. After discharge on local or national laboratories for confirmation. As a confir- May 20, the Daejeon index case-patient felt feverish and matory test, a real-time reverse transcription PCR of nucle- had chills, cough, general weakness, and diarrhea. Because ic acid extracted from sputum specimens was performed of these symptoms, he was hospitalized in hospital A in (5). Cycle threshold values were also measured to quantify Daejeon during May 22–28; the room was shared by 3

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 899 RESEARCH

Table 1. Quarantine policy to prevent additional transmission of MERS, Daejeon, South Korea* Action  The cohort quarantine applied to admitted patients and their caregivers (professional or family) exposed to the MERS case-patients.  Inpatients admitted to the same hospital room before quarantine were quarantined in the same room because their degree of exposure was probably the same. Their caregivers were also quarantined in the same room because of the need for caregiving.  The medical staff (physicians, nurses, and medical technologists) exposed to the MERS case-patients were subjected to home quarantine. However, members of the households of medical staff were not subjected to home quarantine until and unless that medical staff member exhibited any symptoms. Contact between household members and the medical staff member was severely restricted.  The wards under cohort quarantine were controlled by unexposed medical staff using level D protectors (Microguard 2000; 3M, Bracknell, UK). Each protector included an N95 mask, protective glasses, a whole-body protective gown, gloves, and boots.  The body temperature of persons (including inpatients and caregivers) and medical staff admitted to cohort or home quarantine was checked, and these persons were clinically interviewed twice daily. If they reported any symptoms (including a febrile sensation or chills) or if they were asymptomatic but with a body temperature >37.5C°, they were immediately placed in a quarantined area at each hospital. The KCDC performed laboratory tests at this stage; the results were available 3 d later. If the doctor in charge strongly suspected MERS, that patient could be transferred, with careful precautions, to a national isolation hospital within 1 d.  All wards were disinfected by use of sodium hydrosulfite, 80% (vol/vol) alcohol, and 2% (vol/vol) chlorhexidine twice during each shift, thus 6 times/d.  South Korea operates a nationwide medical insurance scheme; all costs incurred by MERS patients were covered.  Persons with confirmed MERS were transferred to another quarantine room that had negative-pressure equipment. Strategies for caregivers  The infection control team carefully explained the risk for MERS and the need for cohort quarantine to all caregivers. Some caregivers did not wish to remain in hospital wards with inpatients. They were taken home and placed in in-home quarantine and used the same MERS quarantine strategy applicable to medical staff in close contact with the patients.  Caregivers attended only noninfected inpatients who required total care. If an inpatient was confirmed to have MERS, nursing care was provided by professional nurses wearing protectors.  The infection control team continuously educated caregivers on how MERS was transmitted and how to prevent infection. Caregivers were told to wear protectors (N95 masks, vinyl gowns, and gloves) and to not touch each other. However, during the first week of quarantine, checks of closed-circuit television footage showed that the protector and contact rules were sometimes not obeyed in hospital A.  Hospital A designated 2 rooms for caregivers in the quarantine ward. The caregivers could use these rooms when they were not actively engaged in patient care. *KCDC, Korea Centers for Disease Control and Prevention; MERS, Middle East respiratory syndrome. inpatients and 1 caregiver. Because his symptoms did not Technical Appendix, https://wwwnc.cdc.gov/EID/article/ improve, he was transferred to the emergency department 23/6/16-0120-Techapp1.pdf). Other than the index case, 14 at hospital B. After hospitalization in ward 101 in hospital cases occurred in hospital A and 11 in hospital B. Case- B, he was suspected of having MERS and was isolated in a patients 30, 38, and 128 were admitted to the same room negative-pressure room on May 30. Ultimately, he became in hospital A as the Daejeon index case-patient. Case-pa- the 16th confirmed MERS patient of 186 total case-patients tient 85 was a caregiver hired by case-patient 128, so she during the 2015 outbreak. was in the same hospital room in hospital A during May Before the Daejeon case-patient was isolated, those 22–28. Case-patients 23, 24, 31, 36, and 95 were admit- around him did not use protective equipment. Therefore, ted to the same room in hospital B as the Daejeon index virus was spread from him during his first 10 days of illness, case-patient. Case-patient 82 was the wife of case-patient before MERS diagnosis and isolation. When we checked 36; case-patient 106 was a caregiver hired by case-patient the closed-circuit television recordings from hospital A to 36; and case-patient 127 was the wife of case-patient 24. estimate how many persons could have been in contact with Therefore, when caregiving, they were in the same room as the Daejeon index case-patient, we found that he had been the Daejeon index case-patient. in several sections of the hospital ward, in particular those The median age of the case-patients was 71 (inter- located on the left side of the nurse station. These sections quartile range 38–86) years; 13 (52.0%) were male; 6 included his admission room, a restroom, the nurse station, (24.0%) were commercial caregivers; and 3 (12.0%) were the foyer, and the hall in front of the elevators. The Daejeon family caregivers. A total of 18 (72.0%) case-patients had index case-patient had potentially contacted every inpatient underlying diseases; 7 (28.0%) had pulmonary diseases, in the same hospital ward. Therefore, we classified all pa- such as asthma, chronic obstructive pulmonary disease, tients and caregivers in that ward as possible contacts. idiopathic pulmonary disease, lung cancer, and pulmonary tuberculosis. Description of Patients with Confirmed Cases All patients reported fever. Other signs and symptoms A total of 26 cases (including the index case) were included chills (10 patients, 38.5%), cough (8, 30.8%), spu- confirmed in the 2 hospitals, and 11 case-patients died tum (6, 23.1%), myalgia (9, 34.6%), headache (4, 15.4%), of MERS (4 in hospital A and 7 in hospital B) (online dyspnea (6, 23.1%), nausea (3, 11.5%), diarrhea (6, 23.1%),

900 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Hospital Outbreaks of MERS, South Korea, 2015

Figure 1. Epidemic curves for the Middle East respiratory syndrome outbreak in Daejeon, South Korea, 2015. The cases are numbered in the order in which they were confirmed in the context of all cases reported during the outbreak. A) Hospitals A and B; B) Hospital A; C) Hospital B. Case-patient 38 is not included because date of illness onset is unknown. Black, weekday; blue, Saturday; red, Sunday or holiday. sore throat (3, 11.5%), rhinorrhea (2, 7.7%), hemoptysis (1, relatively sporadic pattern compared with that for hospital 3.8%), and abdominal discomfort (1, 3.8%) B. The peak in hospital B comprised mostly patients who shared a hospital room with the index case-patient. Most Quarantine MERS cases appeared later in professional or family care- To prevent the spread of MERS-CoV to the local com- givers rather than in inpatients. munity, on June 1, 2015, the government of South Korea The estimated median incubation period for confirmed ordered cohort quarantine, which hospitals A and B fol- cases was 6.1 (95% CI 4.7–7.5) days (Figure 2). Incubation lowed (Table 1). Persons with a history of exposure to periods were 8.8 (95% CI 7.2–10.4) days for hospital A and patients with confirmed MERS were isolated in the same 4.6 (95% CI 2.9–6.2) days for hospital B. hospital ward. In hospital A, the index case-patient was admitted to room 5101, in sector A (Figure 3). Thereafter, 12 case- Epidemic Curve patients were in sector A, and 1 was in sector B. However, After the index case-patient in Daejeon spread MERS-CoV the case-patient in sector B had a history of contact with in Daejeon, the first case occurred on May 30, 2015, and case-patient 85, who was transferred to sector B from sec- the last on June 15, 2015 (total outbreak duration 17 days) tor A for quarantine. Most case-patients were presumed (Figure 1). The epidemic curve for hospital A suggested a to have been infected by the Daejeon index case-patient

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 901 RESEARCH

Figure 2. Estimated incubation periods for the Middle East respiratory syndrome outbreak in Daejeon, South Korea, 2015. Curves indicate estimated cumulative fractions of cases corresponding to the incubation periods, estimated by creating log- normal density functions fitting the observed data. Horizontal lines indicate 95% CIs for the 5th, 50th, and 95th percentiles of the estimated incubation periods. A) Total; estimated median incubation period was 6.1 (95% CI 4.7–7.5) days. B) Hospital A; estimated median incubation period was 8.8 (95% CI 7.2–10.4) days. C) Hospital B; estimated median incubation period was 4.6 (95% CI 2.9–6.2) days.

(case-patient 16). However, 3 instances of other transmis- Discussion sion were noted: case-patient 85 to case-patient 130, case- During the MERS outbreak in South Korea, 25 confirmed patient 54 to case-patient 172, and several case-patients to cases (including 11 deaths) occurred in Daejeon, all associ- case-patient 129. For this last instance of transmission, we ated with the same index case-patient. Two hospitals were could not identify the most probable source, because many affected. The incubation periods and case-fatality rates for possible exposures were evident (case-patients 54, 84, 86, the 2 hospitals differed. 87, 107, and 149). Under the South Korea healthcare system, patients can In hospital B, the index case-patient was admitted to visit secondary hospitals and the emergency rooms of ter- room 1007, located on the upper side of ward 101 (sector tiary hospitals without limitation (5), which probably fa- C). Eight case-patients were in sector C. Case-patient 83 cilitated nosocomial transmission of MERS-CoV. Indeed, was in room 1013 on the opposite side of ward 101 (sector the outbreak in Daejeon accounted for 1 of the 3 major D). Case-patient 45 was in the emergency room and ward MERS-CoV outbreaks in South Korea. These observations 101 with the index case-patient. Case-patient 148 was pre- underscored the importance of the outbreak in Daejeon, sumed to have been infected by case-patient 36 during per- leading the South Korea government to focus resources on formance of cardiopulmonary resuscitation in the intensive controlling transmission of the virus. care unit. The estimated median incubation period for MERS during the outbreak we report was similar to that for out- Attack Rate and Case-Fatality Rate breaks in the eastern province of Saudi Arabia in 2013 A total of 14 cases developed among 301 persons exposed (1). Incubation period estimates may differ, depending in hospital A (attack rate 4.7%) and 11 among 371 persons on the method used to select exposure: the most probable exposed in hospital B (attack rate 3.0%) (Table 2). The at- exposure versus overlapping exposures. In our study, the tack rates for the sectors hosting the index case-patients incubation periods estimated by using both methods were (sector A of hospital A, sector C of hospital B; Figure 3) similar. The incubation period estimated by using the were higher than those for other sectors (sector B of hos- most probable exposure method was 6.1 (95% CI 4.7–7.5) pital A, sector D of hospital B; Figure 3) (31.6% vs. 2.4% days, and that estimated by using the overlapping expo- in hospital A, p<0.05; 18.2% vs. 6.5% in hospital B; Table sures method was 5.6 (95% CI 4.2–6.9) days. The over- 3). The probability of infection for a person admitted to the all attack rate among all exposed persons in Daejeon was same rooms as the index case-patient was 75.0%. In both similar to that for Pyeongtaek (5). The case-fatality rate of hospitals, attack rates were somewhat higher for caregiv- the outbreak in Daejeon was lower than that in the eastern ers (22.5%) than for inpatients (12.3%), although statistical province of Saudi Arabia in 2013 (1) but similar to that in significance was not attained. Jeddah, Saudi Arabia, in 2014 (3). The overall case-fatality rate was 44% (Table 4). This Our results indicate various epidemiologic charac- rate was higher for patients in hospital B (63.6%) than for teristics of MERS-CoV. All persons acquired infection those in hospital A (28.6%), although statistical signifi- in a hospital setting, which is consistent with the previous cance was not attained. outbreak in Saudi Arabia, in which nosocomial spread

902 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Hospital Outbreaks of MERS, South Korea, 2015

Figure 3. Locations of Middle East respiratory syndrome case-patients in hospitals A and B, Daejeon, South Korea, 2015, showing where case- patients were exposed to presumed infectors. Not shown are case- patient 143, an engineer working in hospital A, because the location of his exposure is unclear; case-patient 45, a family caregiver in either the emergency department or room 1015; and case-patient 148, a nurse in the intensive care unit.

was a major route of MERS-CoV transmission (1). The on the fifth floor are occupied by bedridden patients. The characteristics of the specific hospital seemed to affect attack rate among caregivers was higher in hospital A attack rates and case-fatality rates. The index case-pa- than in hospital B. Because immobile patients require tient in Daejeon was consecutively admitted to 2 hos- personal caregiving, their caregivers were required to be pitals. The fifth floor of hospital A specializes in senile in prolonged close contact with patients, which might patients, most of whom have chronic illnesses, includ- have resulted in a higher attack rate. Hospital B is a uni- ing Parkinson’s disease, paraplegia attributable to old versity hospital and thus contained more severely ill pa- infarctions, or amyotrophic lateral sclerosis. Most beds tients than hospital A. Ward 101, to which the Daejeon

Table 2. Middle East respiratory syndrome attack rates among all exposed persons, Daejeon, South Korea, 2015 Hospital A Hospital B Total No. exposed /no. Attack No. exposed/no. Attack No. exposed/no. Attack Exposure with confirmed case rate, % with confirmed case rate, % with confirmed case rate, % Total 301/14 4.7 371/11 3.0 672/25 3.7 Inpatients 227/8 3.5 122/6 4.9 349/14 4.0 Same ward as index case- 62/8 12.9* 52/6 11.5* 114/14 12.3* patient Other wards 165/0 0 70/0 0 235/0 0 Caregivers† 29/5 17.2* 32/4 12.5 61/9 14.8* Same ward as index case- 17/5 29.4‡ 23/4 17.4 40/9 22.5* patient Other wards 12/0 0 9/0 0 21/0 0 Nurses 20/0 0 78/1 1.3 98/1 1.0 Doctors 8/0 0 35/0 0 43/0 0 Others§ 17/1 5.9 104/0 0 121/1 0.8 *p<0.05 †Family or professional. ‡p<0.1. §Paramedics, students, engineers, and visitors.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 903

RESEARCH

Table 3. Attack rates for Middle East respiratory syndrome among inpatients and caregivers in the same wards as the index case- patient, Daejeon, South Korea, 2015 Hospital A Hospital B Total No. exposed/no. with Attack No. exposed/no. Attack No. exposed/no. Attack Person confirmed case rate, % with confirmed case rate, % with confirmed case rate, % Total 79/13 16.5 75/10 13.3 154/23 14.9 Sex M 20/5 25.0 36/7 19.4 56/12 21.4* F 59/8 13.6 39/3 7.7 98/11 11.2 Age, y 30–64 28/5 17.9 28/3 10.7 56/8 14.3 >65 51/8 15.7 47/7 14.9 98/15 15.3 Role Inpatient 62/8 12.9 52/6 11.5 114/14 12.3 Caregiver† 17/5 29.4 23/4 17.4 40/9 22.5 Hospital or room Same sector‡ 38/12 31.6§ 44/8 18.2 82/20 24.4§ Same room 4/4 100¶ 12/8 66.7¶ 16/12 75.0¶ Other room 34/8 23.5 32/0 0 66/8 12.1 Different sector# 41/1 2.4 31/2 6.5 72/3 4.2 Ambulatory Yes 32/10 31.3 41/7 17.1 73/17 23.3 No 23/3 13.0 13/3 23.1 36/6 16.7 Not known 24/0 0 21/0 0 45/0 0 *p<0.1. †Family or professional. ‡Sector A of hospital A or sector C of hospital B, as described in Figure 3. §p<0.05 when compared with attack rates in a different sector. ¶p<0.05 when compared with attack rates in other rooms. #Sector B of hospital A or sector D of hospital B (Figure 3).

index case-patient was admitted, is the main ward of the the same hospital room; thus, if 1 person in the cohort was pulmonary medicine department. We presumed that the infected by MERS-CoV, others were exposed, increasing case-fatality rate was higher for hospital B than hospital the probability of MERS-CoV transmission. This practice A because of underlying pulmonary disease, which has raises an ethical issue in terms of whether letting persons been reported to be a risk factor for development of more stay in the same room with potential MERS patients is severe diseases (8). justified by the purpose of preventing transmission of the Generally, cohort quarantine may be useful as an virus to the community. Some caregivers at hospital A infection-control tool to limit virus transmission in hos- may have had difficulty complying with the quarantine pitals in which susceptible inpatients are gathered or to policy (the protector and contact rules) because they cared more effectively detect infected patients 9( ). Hospitals A for immobile patients. Thus, this practice may have in- and B applied cohort quarantine. In this situation, cohort creased transmission within the hospital. quarantine had several advantages and disadvantages. We identified several cases with uncommon routes The primary purpose was to prevent the spread of MERS- of transmission. Case-patient 148 was the head nurse CoV to the local community. After applying cohort quar- of the intensive care unit to which case-patient 36 was antine, no further spread of MERS-CoV occurred other admitted. When case-patient 36 experienced cardiac ar- than within hospitals A and B. This result may have been rest, that nurse performed cardiopulmonary resuscitation achieved by quarantining all persons who had been in while wearing a level D protector. However, afterward, contact with MERS-CoV–infected patients and by refus- she may have been exposed to MERS-CoV when she ing hospital entry to all susceptible persons. In addition, wiped sweat with her bare arm. Case-patient 143 was an more cases were diagnosed promptly by active surveil- employee of hospital A; he worked in information tech- lance of the cohort. However, this policy had a limitation. nology. He was employed by hospital A during Janu-

One cohort accommodated inpatients and caregivers in ary–May 30, 2015, and his bedroom was located on the

Table 4. Case-fatality rates among all Middle East respiratory syndrome case-patients, Daejeon, South Korea, 2015 Hospital A Hospital B Total Concurrent No. incident cases/ Case-fatality No. incident cases/ Case-fatality No. incident cases/ Case-fatality condition no. fatal cases rate, % no. fatal cases rate, % no. fatal cases rate, % Total 14/4 28.6 11/7 63.6 25/11 44.0 Pulmonary 1/0 0 6/5 83.3 7/5 71.4 None or 13/4 30.8 5/2 40 18/6 33.3 nonpulmonary

904 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Hospital Outbreaks of MERS, South Korea, 2015 seventh floor. His routine work routes, shown on closed- References circuit television, did not reveal any close contact with 1. Assiri A, McGeer A, Perl TM, Price CS, Al Rabeeah AA, Cummings DA, et al.; KSA MERS-CoV Investigation Team. case-patients; thus, the transmission route was unclear. Hospital outbreak of Middle East respiratory syndrome coronavi- We presume that he was infected by fomites in an eleva- rus. N Engl J Med. 2013;369:407–16. http://dx.doi.org/10.1056/ tor or exposed to a patient in a place lacking closed- NEJMoa1306742 circuit television coverage. 2. Al-Tawfiq JA, Memish ZA. Middle East respiratory syndrome coronavirus: epidemiology and disease control measures. Infect The transmission route for case-patient 83 was also Drug Resist. 2014;7:281–7. unidentified. It is possible that some medical staff and 3. Oboho IK, Tomczyk SM, Al-Asmari AM, Banjar AA, Al-Mugti H, caregiver, contaminated with MERS-CoV after visiting Aloraini MS, et al. 2014 MERS-CoV outbreak in Jeddah—a link to room 1007, may have visited case-patient 83 in room health care facilities. N Engl J Med. 2015;372:846–54. http://dx.doi.org/10.1056/NEJMoa1408636 1013. Of note, when case-patient 83 was exposed to 4. Kim SW, Yang TU, Jeong Y, Park JW, Lee K-J, Kim KM, et al. MERS-CoV, the outbreak in Daejeon had not yet been Middle East respiratory syndrome coronavirus outbreak in the recognized and hospital B had not yet implemented in- Republic of Korea. Osong Public Health and Research Perspectives. fection control strategies (e.g., handwashing; wearing 2015;6:269–78. http://dx.doi.org/10.1016/j.phrp.2015.08.006 5. Ki M. 2015 MERS outbreak in Korea: hospital-to-hospital gloves, masks, and vinyl gowns). transmission. Epidemiol Health. 2015;37:e2015033. This study had several limitations. First, we cannot http://dx.doi.org/10.4178/epih/e2015033 be certain that all chains of infection between case-pa- 6. Kim KM, Ki M, Cho SI, Sung M, Hong JK, Cheong HK, et al.. tients have been identified. We did not perform serologic Epidemiologic features of the first MERS outbreak in Korea: focus on Pyeongtaek St. Mary’s Hospital. Epidemiol Health. analyses to seek cases that were potentially missed; such 2015;37:e2015041. http://dx.doi.org/10.4178/epih/e2015041 missed cases may have affected the incubation period esti- 7. Reich NG, Lessler J, Cummings DA, Brookmeyer R. mates and case-fatality rate. Second, closed-circuit televi- Estimating incubation period distributions with coarse data. sion may not have captured all relevant movements. Stat Med. 2009;28:2769–84. http://dx.doi.org/10.1002/sim.3659 8. Assiri A, Al-Tawfiq JA, Al-Rabeeah AA, Al-Rabiah FA, In conclusion, in 2015, Daejeon experienced a hos- Al-Hajjar S, Al-Barrak A, et al. Epidemiological, demographic, pital-associated outbreak of MERS-CoV. Two hospitals and clinical characteristics of 47 cases of Middle East respiratory experienced nosocomial outbreaks, and virus transmission syndrome coronavirus disease from Saudi Arabia: a descriptive was evident among mostly inpatients and caregivers. To study. Lancet Infect Dis. 2013;13:752–61. http://dx.doi.org/ 10.1016/S1473-3099(13)70204-4 prevent the spread of the virus to the local community, 9. Reynolds DL, Garay JR, Deamond SL, Moran MK, Gold W, we developed a unique and successful cohort quarantine Styra R. Understanding, compliance and psychological impact of policy. However, ethical issues associated with this policy the SARS quarantine experience. Epidemiol Infect. 2008;136:997– require thorough discussion by policy makers. 1007. http://dx.doi.org/10.1017/S0950268807009156

Dr. Park is a medical specialist in infectious diseases and an Address for correspondence: Hae-Sung Nam, Department of Preventive epidemiologist at the Korea Centers for Disease Control and Medicine and Public Health, Chungnam National University School of Prevention, in Cheongju, South Korea. His research interest is Medicine, 266 Munhwa-ro, Jung-gu, Daejeon, 301-747, South Korea; MERS-CoV infection. email: [email protected]

EID Podcast: Unraveling the Mysteries of Middle East Respiratory Syndrome Coronavirus Middle East respiratory syndrome coronavirus (MERS-CoV) is a novel CoV known to cause severe acute respiratory illness in humans; approximately 40% of confirmed cases have been fatal. Human-to-human transmission and multiple outbreaks of respiratory illness have been attributed to MERS-CoV, and severe respiratory illness caused by this virus continues to be identified. As of February 23, 2014, the World Health Organization has reported 182 laboratory-confirmed cases of MERS-CoV infection, including 79 deaths, indicating an ongoing risk for transmission to humans in the Arabian Peninsula. Visit our website to listen: https://www2c.cdc.gov/podcasts/player.asp?f=8631627

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 905 RESEARCH

Genomic Analysis of Salmonella enterica Serovar Typhimurium DT160 Associated with a 14-Year Outbreak, New Zealand, 1998–2012 Samuel J. Bloomfield, Jackie Benschop, Patrick J. Biggs,1 Jonathan C. Marshall,1 David T.S. Hayman,1 Philip E. Carter,1 Anne C. Midwinter, Alison E. Mather, Nigel P. French

During 1998–2012, an extended outbreak of Salmonella en- humans peaks in the warm summer months, probably in terica serovar Typhimurium definitive type 160 (DT160) af- association with increased multiplication of Salmonella in fected >3,000 humans and killed wild birds in New Zealand. animal and food sources and with increased participation in However, the relationship between DT160 within these 2 host higher risk outdoor activities (e.g., activities that increase groups and the origin of the outbreak are unknown. Whole- contact with wild-life) (5). Climate change is expected to genome sequencing was used to compare 109 Salmonella increase summer temperatures, potentially increasing sal- Typhimurium DT160 isolates from sources throughout New Zealand. We provide evidence that DT160 was introduced monellosis incidence in New Zealand (6). into New Zealand around 1997 and rapidly propagated During 1998–2012, an extended outbreak of Salmo- throughout the country, becoming more genetically diverse nella Typhimurium definitive type 160 (DT160) occurred over time. The genetic heterogeneity was evenly distributed in New Zealand (7). During the outbreak, DT160 was the across multiple predicted functional protein groups, and we predominant Salmonella spp. subtype isolated from hu- found no evidence of host group differentiation between man salmonellosis patients and sick wild birds. DT160 isolates collected from human, poultry, bovid, and wild bird was also isolated from other animals and the environment, sources, indicating ongoing transmission between these host but it was not the main Salmonella subtype isolated from groups. Our findings demonstrate how a comparative ge- these sources (8–10). DT160 has been isolated from ani- nomic approach can be used to gain insight into outbreaks, mals and environments worldwide (11,12) and is usually disease transmission, and the evolution of a multihost patho- associated with moribund birds (13,14). However, before gen after a probable point-source introduction. the 1998–2012 outbreak, DT160 had not been reported in New Zealand. In 2009, an outbreak of DT160 involving ontyphoidal serovars of Salmonella enterica subsp. humans and wild birds was reported in Tasmania, Austra- Nenterica, which cause salmonellosis, are responsible lia (15); however, as with the outbreak in New Zealand, for an estimated 93.8 million illnesses and 155,000 deaths the relationship between DT160 within the bird and human among humans worldwide each year (1). In New Zealand, host groups of Tasmania was unknown. We used genomic these serovars are the second largest cause of bacterial gas- epidemiologic approaches to characterize the origin, evolu- troenteritis, annually causing 21 cases per 100,000 population, and transmission of Salmonella Typhimurium DT160 tion (2). Nontyphoidal Salmonella spp. strains vary in host in New Zealand. specificity and are usually transmitted to humans via direct contact or consumption of foods originating from animals Methods (3,4). In New Zealand, salmonellosis incidence among Whole-Genome Sequencing Author affiliations: Massey University, Palmerston North, After stratifying the Salmonella strain collection at the En- New Zealand (S.J. Bloomfield, J. Benschop, P.J. Biggs, teric Reference Laboratory of the Institute of Environmen- J.C. Marshall, D.T.S. Hayman, A.C. Midwinter, N.P. French); tal Science and Research Ltd. (Wallaceville, New Zealand) Institute of Environmental Science and Research, Wellington, by age and host, we randomly selected 35 human, 25 wild New Zealand (P.E. Carter); University of Cambridge, Cambridge, bird, 25 poultry, and 24 bovine DT160 isolates from 1998– UK (A.E. Mather) 2012. We extracted genomic DNA from these isolates DOI: https://dx.doi.org/10.3201/eid2306.161934 1These authors contributed equally to this article.

906 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 S. enterica Associated with 14-Year Outbreak using a QIAamp DNA Mini Kit (QIAGEN, Hilden, Ger- isolate. The position of the SNP on the reference genome many) (16). New Zealand Genomics Limited (NZGL) at was used to determine if both methods identified the SNP Massey Genome Service, Massey University, Palmerston or if they were unique to the method (online Technical Ap- North, New Zealand, performed whole-genome sequencing pendix, https://wwwnc.cdc.gov/EID/article/23/6/16-1934- of the extracts. NZGL also prepared a library for each iso- Techapp1.pdf). This method identified 793 core SNPs late by using a TruSeq DNA PCR-Free Library Preparation shared by the 109 New Zealand DT160 isolates. Kit (Illumina, Scorsby, Victoria, Australia) and sequenced the libraries by using MiSeq (Illumina, San Diego, CA, Global DT160 Strains USA) as 2 × 250 bp paired-end runs (≈120–150 genome Using the genomic assembly and SNP identification meth- coverage). After sequencing and standard barcode demulti- ods as we described, we compared 2 DT160 strains from plexing, NZGL used FASTQ-MCF (17) to perform quality the United Kingdom with the 109 DT160 isolates from control procedures to remove any PhiX control library reads New Zealand: 1,521 core SNPs were identified. We down- and adaptor sequences. The raw reads for the 109 DT160 loaded the UK strains, which were previously published by isolates are available in the European Nucleotide Archive Petrovska et al. (24), from the European Nucleotide Ar- (http://www.ebi.ac.uk/ena; accession no. PRJEB18077). chive (accession nos. ERS015626 and ERS015627).

Genomic Assembly Phylogenetic Inference and Distances Each isolate’s genome was assembled de novo. We used We used RAxML version 8.2.4 (25) to construct a max- an in-house Perl script to trim reads at an error probabil- imum-likelihood tree based on the 793 core SNPs of the ity of 0.01 and generate random subsets of paired reads 109 DT160 isolates; we used EvolView version 2 (26) to from 750,000 to 1.2 million paired reads in increments of visualize and edit the tree. We used SplitsTree (27) to form 150,000, varying the average coverage. We assembled each a NeighborNet tree of the 109 New Zealand DT160 isolates of the random sets by using the de novo assembler Velvet based on the 793 core SNPs that they share and to com- version 1.1 (18) at a variety of k-mers (from 55 to 245) in pare the New Zealand and UK isolates based on the 1,521 increments of 10. De novo assembly resulted in multiple core SNPs that they share. We used MEGA6 (28) and the genome assemblies for each isolate. We ranked the metrics maximum composite likelihood model (29) to predict the for each of 4 parameters (longest genome length, fewest pairwise distance between the 109 New Zealand DT160 number of contigs, largest N50 value, and longest contig isolates, based on the 793 core SNPs they share, and the length) in numeric order and calculated an overall equally 109 New Zealand and 2 UK isolates, based on the 1,521 summed ranking score for each assembly. We used the as- core SNPs that they share. semblies with the lowest total rank for further analyses. We used QUAST (19), a quality assessment tool for evaluating Phylogenetic Analysis and comparing genome assemblies, to analyze the DT160 We used an in-house Perl script to split the 793 codons into de novo assemblies and determine their GC content (i.e., 5 groups: those associated with the first, second, or third the percentage of a DNA sequence made up of guanine and codon; those contained in overlapping coding regions; cytosine bases). and those found in intergenic regions. We also used the in-house Perl script to determine whether the SNPs were Single-Nucleotide Polymorphism Identification synonymous or nonsynonymous. We then exported the We used Snippy version 2.6 (https://github.com/tseemann/ partitioned SNPs into BEAUti to create an XML file for snippy) and kSNP version 3.0 (20) to identify core single- BEAST 1.8.3 (30). nucleotide polymorphisms (SNPs). Snippy is a pipeline To allow for variation in base substitution among codon that uses the Burrows-Wheelers Aligner (21) and SAM- positions, we used separate Hasegawa Kishino Yano models tools version 1.3.1 (22) to align reads from different iso- to estimate the 5 SNP groups (31); to allow for and estimate lates to a sequence and uses FreeBayes (23) to identify changes in the effective population size, we used the Gauss- variants among the alignments. We used kSNP to analyze ian Markov random field Bayesian skyride model 32( ); to al- de novo assembled genomes, along with the reference ge- low for variation in mutation rates among lineages, we used nome, S. enterica serovar Typhimurium 14028S (GenBank an uncorrelated relaxed molecular clock (33), which was accession no. NC_016856). We used an in-house Python calibrated by the tip dates. We ran the XML file in BEAST script to determine the read coverage of all the SNPs iden- for 40 million steps a total of 3 times with different start- tified via kSNP. We used Snippy to align reads from each ing seeds before using LogCombiner (http://beast.bio.ed.ac. isolate to the reference genome (GenBank accession no. uk/LogCombiner) to combine the runs with a 10% burn-in. NC_016856) before identifying SNPs. SNPs were accepted To visualize the results and the relative change in effective if they had a >10 read depth and a >90% consensus for each population size, we used Tracer version 1.6 (34).

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 907 RESEARCH

To determine the mutation rate for the DT160 genome, periods, we applied PERMANOVA (http://www.primer-e. we multiplied the mutation rate estimated by BEAST by com/permanova.htm). To assess differences in dispersions the number of analyzed core SNPs (793 bp) and then di- between different groups, we computed dispersions (z-val- vided the product by the mean genome size of the analyzed ues) by using PermDisp (42) and then modeled them using isolates (4,884,485 bp). We used the discrete phylogeo- a regression model with date of collection and source as the graphic model (35) to predict ancestral migrations between explanatory variables. host groups over the course of the outbreak. Scripts Protein Coding Gene Analysis The in-house scripts used for genomic analyses in this study We used Prokka (36) to annotate de novo assembled ge- were specifically designed for this dataset. The scripts are nomes, and we used Roary (37) to cluster proteins and available from GitHub (https://github.com/samuelbloom- identify those that were found only in a subset of isolates field/Scripts-for-genomic-analyses). and those that differed in length between the isolates. We used ClustalW version 2.1 (38) to align amino acid se- Results quences, and we used an in-house Perl script to determine During the 1998–2012 human outbreak of Salmonella if these alignments contained mismatches. The nucleotide Typhimurium DT160 in New Zealand, disease incidence sequence of all proteins that differed were extracted from displayed a typical epidemic curve: prevalence increased the assembled genomes, along with 500-bp flanks on ei- from 1999 to 2000, before peaking at 791 cases in 2001, ther side of the sequence, by using an in-house Perl script. and then slowly decreased from 2002 through 2012 (Fig- We could not obtain 500-bp flanks for some genes because ure 1). At the same time, numerous isolates were reported they were located at the end of contigs. For those genes, the from nonhuman hosts (wild birds, poultry, bovids), and flank was cut short, but their length was annotated. We ex- disease incidence among these host groups displayed epi- tracted flanks to help with read alignment. This extraction demic curves similar to those for humans (online Techni- left a pool of nucleotide sequences from each isolate, for cal Appendix). every protein that potentially differed in sequence. For each protein, we extracted all nucleotide variants from the pool Genomic DT160 Comparison by using an in-house Perl script. We used SRST2 version The genomes we assembled were 4.8–4.9 Mb in length and 2, a read mapping–based tool (39), to align reads from each had a GC content of 52.11%–52.16% (reference value for isolate to the sequence variants, and we used SAMtools S. enterica 50%–53%) (43). We identified 793 core SNPs version 1.3.1 (22) to form a consensus sequence from the shared by the 109 DT160 isolates from New Zealand. aligned reads. We set the consensus cutoff at a read depth of >8 and a consensus of >80%. The flanks were removed from the consensus sequences, and the sequence variants were translated into amino acid sequences by using an in-house Perl script. We identified protein differences by comparing the amino acid sequences from each isolate and combined the differences with the nonsynonymous SNPs identified by SNP analysis. The position of nonsynonymous SNPs within proteins was used to prevent repeats. We used the Clusters of Orthologous Groups of proteins (COGs) database (40) to predict protein functions. For each functional group, we calculated the proportion of proteins that differed in sequence, and we used a Fisher exact test, computed via Monte Carlo Markov Chains of ≈109 it- erations, to determine if there were any differences between these proportions. We used an in-house Perl script to form a presence–absence matrix of all the protein differences. We used Primer- E version 6 (41) to predict the Euclidian distance between the isolates based on the presence–absence matrix. The centroid is the arithmetic mean for a group of data points in Figure 1. Number of Salmonella enterica serovar Typhimurium an n-dimensional space. To assess differences in centroids DT160 cases and isolates reported during an outbreak in New Zealand, 1998–2012. A) Cases in humans (8,9). B) Isolates from among isolates collected from different sources or time nonhuman sources (8,10).

908 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 S. enterica Associated with 14-Year Outbreak

DT160 Introduction Date Salmonella Typhimurium strains in sub-Saharan Africa. Ancestral date reconstruction analysis predicted that the The similarity of these mutation rates suggests consisten- 109 New Zealand DT160 isolates shared a date of common cy between outbreaks caused by S. enterica serovar Ty- ancestor in approximately August 1997 (95% highest pos- phimurium and has implications for modeling the evolution terior density interval June 1996–August 1998). Compara- of future outbreaks caused by this serovar. tive analysis indicated that the 2 DT160 isolates collected In bacteriology, the effective population size is the from the United Kingdom were genetically distinct from number of bacteria that contribute to the next generation. the 109 New Zealand DT160 isolates (online Technical The increase in the DT160 effective population size during Appendix). The average pairwise SNP distance between 1998–2003 coincided with an increased prevalence of DT160 the 2 UK DT160 isolates and the New Zealand isolates among human and nonhuman hosts during this time. Howev- was 0.0287, compared with an average pairwise distance er, the subsequent levelling-off of the effective DT160 popu- of 0.0151 between New Zealand isolates. lation size is probably an artifact because we calculated the In New Zealand, DT160 was first reported in Christ- effective population size from the timing of coalescent events church in 1998 from a human with salmonellosis (44) (an for randomly sampled bacteria (32), and as the outbreak pro- isolate from this case was included as part of this study). ceeded, fewer coalescent points were available for estimation. The New Zealand DT160 isolates we analyzed were estimated to share a common ancestor 0–2 years before this case and were distinct from the UK isolates analyzed, suggesting that DT160 was probably introduced into New Zea- land as a single incursion within this time period. How- ever, worldwide comparative studies are required to track DT160 migration and validate this hypothesis.

DT160 Evolution Our phylogenetic analysis also predicted that the 109 DT160 isolates mutated at a rate of 3.3–4.3 × 10-7 substitutions/site/year (95% highest posterior density interval) and that the effective population size for DT160 increased from 1998 to 2003 (Figure 2). Over the course of the outbreak, DT160 also increased in genetic diversity (Figure 3). The mutation rate estimated for the DT160 outbreak is similar to rates reported by Mather et al. (45) for an outbreak of Salmonella Typhimurium DT104 in Scotland during 1990–2012 and by Okoro et al. (46) for invasive

Figure 3. A) NeighborNet tree of 109 Salmonella enterica serovar Figure 2. Relative effective population size (log scale) of Salmonella Typhimurium DT160 isolates collected during an outbreak in New enterica serovar Typhimurium DT160 during an outbreak in New Zealand, 1998–2012. The tree was based on 793 core single- Zealand, 1998–2012. Population parameters were estimated using nucleotide polymorphisms. Colors indicate date of isolate collection. the Gaussian Markov random field Bayesian skyride model. The The scale bar represents the number of nucleotide substitutions black line represents the median effective population size estimate; per site. B) Scatterplot of the mean pairwise distance of 106 DT160 gray shading represents the 95% highest posterior density interval. isolates from 2000–2011. Error bars represent 95% CIs.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 909 RESEARCH

Overall, our phylogenetic analyses suggest that the poultry, and wild birds (S. Omar, master’s thesis, 2011; DT160 population increased dramatically in the first few http://mro.massey.ac.nz/handle/10179/2681?show=full); years following introduction. As the DT160 population in- however, PFGE could not distinguish DT160 from the creased, it acquired multiple SNPs, resulting in a progres- separate sources. In our study, we were able to use whole- sive increase in diversity over time. genome sequencing to distinguish DT160 at the isolate level. However, we did not find any distinct DT160 clades DT160 Sources associated with any one source (Figure 4). PFGE (pulsed-field gel electrophoresis) was previously Identifying the source of a salmonellosis outbreak used to compare New Zealand DT160 isolates from humans, can be difficult because multiple potential sources must

Figure 4. Maximum-likelihood tree of 109 Salmonella enterica serovar Typhimurium DT160 isolates collected during an outbreak in New Zealand, 1998–2012. The tree was based on 793 core single-nucleotide polymorphisms. Colored squares to the right of the branches indicate the source of isolates. The scale bar represents number of nucleotide substitutions per site. The heat map represents the Euclidean pairwise distance between isolates (based on the presence of 684 protein differences). Isolates that shared a small number of protein differences contained small Euclidean distances and are closer to blue in color on the heat map; isolates that shared a large number of protein differences contained large Euclidean distances and are closer to red in color. The gray squares represent the 2 outliers missing a large number of genes. The diagonal array of blue squares represents the pairwise distance for the same isolates.

910 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 S. enterica Associated with 14-Year Outbreak be considered (47). Probable sources of Salmonella can deletions). The nonsense SNPs and INDELs were respon- be identified by comparing isolates from infected- hu sible for 123 proteins that differed in length. Overall, we mans with those from other human, nonhuman, and en- identified 684 differences in 604 protein sequences among vironmental sources (48). We did not find distinct DT160 the 109 DT160 isolates. We excluded 2 isolates from pro- clades associated with any 1 source, suggesting that after tein coding gene analysis because they were missing a large its introduction into New Zealand, DT160 was transmit- number of proteins (online Technical Appendix). ted between multiple hosts, resulting in large epidemics By using PERMANOVA, we found that centroids among humans and wild birds. Our results also suggest based on the 684 protein differences were indistinguishable that humans obtained DT160 from multiple sources over among groups of DT160 isolates collected from different the course of the outbreak. This finding is consistent with sources and time periods (online Technical Appendix). that in a case-control study performed by Thornley et al. PERMANOVA’s inability to distinguish centroids appears (44), which found that human DT160 cases were associ- to be due to the fact that DT160 isolates radiated out from ated with multiple risk factors involving different sources: a point source. The z-value is the distance from an isolate handling dead wild birds, contact with persons with diar- to the centroid of a group of isolates; we calculated the z- rhea, and consumption of fast food. value for 107 DT160 isolates on the basis of 684 protein differences. Our regression modeling results showed that Ancestral Migration between Hosts the z-value was associated with the date, but not source, of We used the discrete phylogeographic model to predict collection (Figure 5). ancestral migration of DT160 between the animal and hu- The 684 protein differences shared by the DT160 iso- man host groups, similar to Mather et al. (45). However, we lates were associated with a large number of COG func- were unable to detect a signal that could not be attributed tional groups. The proportion of proteins that contained to different sampling fractions in the host groups (online sequence differences differed between functional groups (p Technical Appendix). Therefore, an alternate method, larg- = 0.00002). The proportions varied from 0.06 to 0.18, al- er sample size, or both are required to predict these ances- though most were between 0.09 and 0.13 (online Technical tral migrations. Appendix). In addition, our data were insufficient to model the effects of source or date of collection on the number Protein and Gene Analysis of protein differences associated with each group (online Protein annotation identified 5,096 coding DNA sequences, Technical Appendix). of which 4,983 (98%) were found in all of the isolates, 108 Bacteria often adapt to new environments by alter- (2%) were found in 95%–99% of isolates, and 3 (<1%) were ing (changing or losing) genes that are not essential for found in 1%–5% of the isolates. Protein coding gene analy- colonizing that environment (49). Gene loss can result in sis also identified 477 nonsynonymous SNPs, of which 27 an increase in bacterial fitness, as fewer genes and pro- were nonsense mutations and 96 were INDELs (insertions/ cesses need to be maintained within the bacteria (50). We

Figure 5. Scatter plots of year of collection versus z-values for 107 Salmonella enterica serovar Typhimurium DT160 isolates collected during an outbreak in New Zealand, 1998–2012. Of the 107 isolates, 25 were from poultry (A), 25 from wild birds (B), 24 from bovids (C), and 33 from humans (D). Black lines represent the regression equation; gray shading represents SE for this equation. Date of collection was significantly associated with z-values in this model (p<2−16). There was insufficient evidence to suggest that source was associated with z-values (p = 0.558), and the interaction between source and date of collection was not significant (p = 0.458).

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 911 RESEARCH identified multiple protein changes among the DT160 5. Lal A, Baker MG, French NP, Dufour M, Hales S. The isolates, and these changes occurred in multiple COG epidemiology of human salmonellosis in New Zealand, 1997–2008. Epidemiol Infect. 2012;140:1685–94. http://dx.doi.org/10.1017/ functional groups as the epidemic progressed. However, S0950268811002470 we found no evidence of host group differentiation, sug- 6. Lal A, Ikeda T, French N, Baker MG, Hales S. Climate variability, gesting that most of the evolution was due to random weather and enteric disease incidence in New Zealand: time series genetic drift rather than adaptive evolution. analysis. PLoS One. 2013;8:e83484. http://dx.doi.org/10.1371/ journal.pone.0083484 7. Alley MR, Connolly JH, Fenwick SG, Mackereth GF, Leyland MJ, Discussion Rogers LE, et al. An epidemic of salmonellosis caused by Using genomic analysis, we described the evolution and Salmonella Typhimurium DT160 in wild birds and humans emergence of Salmonella Typhimurium DT160 within in New Zealand. N Z Vet J. 2002;50:170–6. http://dx.doi.org/ 10.1080/00480169.2002.36306 New Zealand. Our results suggest that DT160 was intro- 8. Institute of Environmental Science and Research Ltd (ESR). duced into New Zealand on a single occasion from 1996 ESR LabLink. Quarterly surveillance summaries for New through 1998, before propagating throughout the country Zealand, March 2000–March 2003 [cited 2016 Nov 25]. and becoming more genetically diverse over time. In addi- https://surv.esr.cri.nz/PDF_surveillance/Lablink/ 9. Institute of Environmental Science and Research Ltd (ESR). Public tion, we found that DT160 isolates collected from human, Health Surveillance; Information for New Zealand public health poultry, bovine and wild bird sources were highly similar, action. 2003–2012 human Salmonella isolates [cited 2016 Nov 25]. indicating a large number of transmission episodes be- https://surv.esr.cri.nz/enteric_reference/human_salmonella.php tween these host groups. 10. Institute of Environmental Science and Research Ltd (ESR). Public Health Surveillance; Information for New Zealand public health action. Non-human Salmonella isolates, 2003–2012 Acknowledgments [cited 2016 Nov 25]. https://surv.esr.cri.nz/enteric_reference/ We thank the staff at the Massey Genome Service (part of nonhuman_salmonella.php New Zealand Genomics Ltd), Massey University, Palmerston 11. Penfold JB, Amery HC, Peet PJ. Gastroenteritis associated with wild birds in a hospital kitchen. Br Med J. 1979;2:802. North, New Zealand, for their help and advice with the genome 12. Tizard IR, Fish NA, Harmeson J. Free flying sparrows as carriers of sequencing for this study; Tim G. Vaughan for his help with salmonellosis. Can Vet J. 1979;20:143–4. phylogenetic analyses; and the Enteric Reference Laboratory, 13. Lawson B, Howard T, Kirkwood JK, Macgregor SK, Perkins M, Environmental Science and Research, Wallaceville, Robinson RA, et al. Epidemiology of salmonellosis in garden birds in England and Wales, 1993 to 2003. EcoHealth. 2010;7:294–306. New Zealand for supplying Salmonella isolates. http://dx.doi.org/10.1007/s10393-010-0349-3 14. Piccirillo A, Mazzariol S, Caliari D, Menandro ML. Salmonella This study was funded by the Allan Wilson Centre for Molecular Typhimurium phage type DT160 infection in two Moluccan Ecology and Evolution, Massey University; the Biotechnology cockatoos (Cacatua moluccensis): clinical presentation and and Biological Sciences Research Council (BB/M014088/1 to pathology. Avian Dis. 2010;54:131–5. http://dx.doi.org/10.1637/ A.E.M); and the Royal Society of New Zealand Marsden Fund 8969-062509-Case.1 15. Grillo T, Post L. Salmonella Typhimurium DT160 outbreak in (MAU1503 to D.T.S.H.). Tasmania. Animal Health Surveillance Quarterly Reports. Mr. Bloomfield is a PhD student at Massey University, 2010;14:8–8. http://www.sciquest.org.nz/node/72986 16. QIAGEN. QIAamp DNA mini and blood mini handbook. Palmerston North, New Zealand. His research interests include Third edition. 2012 [cited 2015 Feb 12]. https://moodle.ufsc.br/ infectious diseases and genomics, particularly how infectious pluginfile.php/1379318/mod_resource/content/0/QIAamp_DNA_ diseases are transmitted and evolve over the course of outbreaks. Mini_Blood.pdf 17. Aronesty E. Comparison of sequencing utility programs. The Open Bioinformatics Journal. 2013;7:1–8. http://dx.doi.org/ References 10.2174/1875036201307010001 1. Majowicz SE, Musto J, Scallan E, Angulo FJ, Kirk M, O’Brien SJ, 18 Zerbino DR, Birney E. Velvet: algorithms for de novo short read et al.; International Collaboration on Enteric Disease ‘Burden of assembly using de Bruijn graphs. Genome Res. 2008;18:821–9. Illness’ Studies. The global burden of nontyphoidal Salmonella http://dx.doi.org/10.1101/gr.074492.107 gastroenteritis. Clin Infect Dis. 2010;50:882–9. http://dx.doi.org/ 19. Gurevich A, Saveliev V, Vyahhi N, Tesler G. QUAST: 10.1086/650733 quality assessment tool for genome assemblies. Bioinformatics. 2. Institute of Environmental Science and Research Ltd. (ESR). 2013;29:1072–5. http://dx.doi.org/10.1093/bioinformatics/btt086 Notifiable diseases. New Zealand public health observatory. 2014 20. Gardner SN, Slezak T, Hall BG. kSNP3.0: SNP detection and [cited 2016 Feb 2]. http://www.nzpho.org.nz/NotifiableDisease.aspx phylogenetic analysis of genomes without genome alignment 3. King N, Lake R, Campbell D. Source attribution of nontyphoid or reference genome. Bioinformatics. 2015;31:2877–8. salmonellosis in New Zealand using outbreak surveillance data. http://dx.doi.org/10.1093/bioinformatics/btv271 J Food Prot. 2011;74:438–45. http://dx.doi.org/10.4315/ 21. Li H, Durbin R. Fast and accurate short read alignment with 0362-028X.JFP-10-323 Burrows-Wheeler transform. Bioinformatics. 2009;25:1754–60. 4. Eswarappa SM, Janice J, Nagarajan AG, Balasundaram SV, http://dx.doi.org/10.1093/bioinformatics/btp324 Karnam G, Dixit NM, et al. Differentially evolved genes of 22. Li H. A statistical framework for SNP calling, mutation discovery, Salmonella pathogenicity islands: insights into the mechanism association mapping and population genetical parameter estimation of host specificity inSalmonella. PLoS One. 2008;3:e3829. from sequencing data. Bioinformatics. 2011;27:2987–93. http://dx.doi.org/10.1371/journal.pone.0003829 http://dx.doi.org/10.1093/bioinformatics/btr509

912 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 S. enterica Associated with 14-Year Outbreak

23. Garrison E, Marth G. Haplotype-based variant detection from matrix choice. Nucleic Acids Res. 1994;22:4673–80. short-read sequencing. 2012 [cited 2017 Feb 2]. https://arxiv.org/ http://dx.doi.org/10.1093/nar/22.22.4673 abs/1207.3907 39. Inouye M, Dashnow H, Raven LA, Schultz MB, Pope BJ, 24. Petrovska L, Mather AE, AbuOun M, Branchu P, Harris SR, Tomita T, et al. SRST2: rapid genomic surveillance for public Connor T, et al. Microevolution of monophasic Salmonella health and hospital microbiology labs. Genome Med. 2014;6:90. Typhimurium during epidemic, United Kingdom, 2005–2010. http://dx.doi.org/10.1186/s13073-014-0090-6 Emerg Infect Dis. 2016;22:617–24. http://dx.doi.org/10.3201/ 40. Tatusov RL, Galperin MY, Natale DA, Koonin EV. The COG eid2204.150531 database: a tool for genome-scale analysis of protein functions 25. Stamatakis A. RAxML version 8: a tool for phylogenetic and evolution. Nucleic Acids Res. 2000;28:33–6. http://dx.doi.org/ analysis and post-analysis of large phylogenies. Bioinformatics. 10.1093/nar/28.1.33 2014;30:1312–3. http://dx.doi.org/10.1093/bioinformatics/btu033 41. Clark KR, Gorle RN. PRIMER v6: user manual/tutorial. Plymouth 26. He Z, Zhang H, Gao S, Lercher MJ, Chen WH, Hu S. Evolview v2: (UK): PRIMER-E; 2006. p. 296. an online visualization and management tool for customized and 42. Anderson MJ. Distance-based tests for homogeneity of multivariate annotated phylogenetic trees. Nucleic Acids Res. 2016;44: dispersions. Biometrics. 2006;62:245–53. http://dx.doi.org/10.1111/ W236-41. http://dx.doi.org/10.1093/nar/gkw370 j.1541-0420.2005.00440.x 27. Huson DH, Bryant D. Application of phylogenetic networks in 43. Popoff MY, Le Minor LE. Genus XXXIII. Salmonella. In: evolutionary studies. Mol Biol Evol. 2006;23:254–67. Brenner DJ, Staley JT, editors. Bergey’s manual of systematic http://dx.doi.org/10.1093/molbev/msj030 bacteriology. New York: Springer; 2005. p. 764–99. 28. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. 44. Thornley CN, Simmons GC, Callaghan ML, Nicol CM, Baker MG, MEGA6: Molecular Evolutionary Genetics Analysis version 6.0. Gilmore KS, et al. First incursion of Salmonella enterica serotype Mol Biol Evol. 2013;30:2725–9. http://dx.doi.org/10.1093/molbev/ Typhimurium DT160 into New Zealand. Emerg Infect Dis. mst197 2003;9:493–5. http://dx.doi.org/10.3201/eid0904.020439 29. Tamura K, Nei M, Kumar S. Prospects for inferring very large 45. Mather AE, Reid SWJ, Maskell DJ, Parkhill J, Fookes MC, phylogenies by using the neighbor-joining method. Proc Natl Harris SR, et al. Distinguishable epidemics of multidrug-resistant Acad Sci U S A. 2004;101:11030–5. http://dx.doi.org/10.1073/ Salmonella Typhimurium DT104 in different hosts. Science. pnas.0404206101 2013;341:1514–7. http://dx.doi.org/10.1126/science.1240578 30. Drummond AJ, Suchard MA, Xie D, Rambaut A. Bayesian 46. Okoro CK, Kingsley RA, Connor TR, Harris SR, Parry CM, phylogenetics with BEAUti and the BEAST 1.7. Mol Biol Evol. Al-Mashhadani MN, et al. Intracontinental spread of human 2012;29:1969–73. http://dx.doi.org/10.1093/molbev/mss075 invasive Salmonella Typhimurium pathovariants in sub-Saharan 31. Hasegawa M, Kishino H, Yano T. Dating of the human–ape Africa. Nat Genet. 2012;44:1215–21. http://dx.doi.org/10.1038/ splitting by a molecular clock of mitochondrial DNA. J Mol Evol. ng.2423 1985;22:160–74. http://dx.doi.org/10.1007/BF02101694 47. Gieraltowski L, Julian E, Pringle J, Macdonald K, Quilliam D, 32. Minin VN, Bloomquist EW, Suchard MA. Smooth skyride Marsden-Haug N, et al. Nationwide outbreak of Salmonella through a rough skyline: Bayesian coalescent-based inference of Montevideo infections associated with contaminated imported population dynamics. Mol Biol Evol. 2008;25:1459–71. black and red pepper: warehouse membership cards provide critical http://dx.doi.org/10.1093/molbev/msn090 clues to identify the source. Epidemiol Infect. 2013;141:1244–52. 33. Drummond AJ, Ho SYW, Phillips MJ, Rambaut A. Relaxed http://dx.doi.org/10.1017/S0950268812001859 phylogenetics and dating with confidence. PLoS Biol. 2006;4:e88. 48. Byrne L, Fisher I, Peters T, Mather A, Thomson N, Rosner B, http://dx.doi.org/10.1371/journal.pbio.0040088 et al.; International Outbreak Control Team. A multi-country 34. Rambaut A, Suchard MA, Xie D, Drummond AJ. Beast. Tracer 1.6 outbreak of Salmonella Newport gastroenteritis in Europe 2014 [cited 2016 Sep 27]. http://beast.bio.ed.ac.uk/Tracer associated with watermelon from Brazil, confirmed by whole 35. Lemey P, Rambaut A, Drummond AJ, Suchard MA. Bayesian genome sequencing: October 2011 to January 2012. Euro phylogeography finds its roots. PLOS Comput Biol. 2009; Surveill. 2014;19: 20866. http://dx.doi.org/10.2807/1560-7917. 5:e1000520. http://dx.doi.org/10.1371/journal.pcbi.1000520 ES2014.19.31.20866 36. Seemann T. Prokka: rapid prokaryotic genome annotation. 49. Hottes AK, Freddolino PL, Khare A, Donnell ZN, Liu JC, Bioinformatics. 2014;30:2068–9. http://dx.doi.org/10.1093/ Tavazoie S. Bacterial adaptation through loss of function. bioinformatics/btu153 PLoS Genet. 2013;9:e1003617. http://dx.doi.org/10.1371/ 37. Page AJ, Cummins CA, Hunt M, Wong VK, Reuter S, Holden MTG, journal.pgen.1003617 et al. Roary: rapid large-scale prokaryote pan genome analysis. 50. Koskiniemi S, Sun S, Berg OG, Andersson DI. Selection- Bioinformatics. 2015;31:3691–3. http://dx.doi.org/10.1093/ driven gene loss in bacteria. PLoS Genet. 2012;8:e1002787. bioinformatics/btv421 http://dx.doi.org/10.1371/journal.pgen.1002787 38. Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through Address for correspondence: Samuel J. Bloomfield, Massey University, sequence weighting, position-specific gap penalties and weight Palmerston North, New Zealand; email: [email protected]

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 913 RESEARCH

Stockpiling Ventilators for Influenza Pandemics Hsin-Chan Huang,1 Ozgur M. Araz, David P. Morton, Gregory P. Johnson, Paul Damien, Bruce Clements, Lauren Ancel Meyers

In preparing for influenza pandemics, public health agen- personal protection equipment. Given uncertainty in the cies stockpile critical medical resources. Determining ap- timing and severity of the next pandemic, as well as the propriate quantities and locations for such resources can time required to manufacture medical countermeasures, be challenging, given the considerable uncertainty in the stockpiling is central to influenza preparedness (3). How- timing and severity of future pandemics. We introduce a ever, difficulty in forecasting and limited public health bud- method for optimizing stockpiles of mechanical ventilators, gets often constrain decisions about sizes, locations, and which are critical for treating hospitalized influenza patients in respiratory failure. As a case study, we consider the US deployment of such stockpiles. state of Texas during mild, moderate, and severe pandem- Mechanical ventilators are essential for treating in- ics. Optimal allocations prioritize local over central storage, fluenza patients in severe acute respiratory failure. Sub- even though the latter can be deployed adaptively, on the stantial concern exists that intensive care units (ICUs) basis of real-time needs. This prioritization stems from high might have insufficient resources to treat all persons -re geographic correlations and the slightly lower treatment quiring ventilator support. Prior studies argue that cur- success assumed for centrally stockpiled ventilators. We rent capacities are insufficient to handle even moderately developed our model and analysis in collaboration with ac- severe pandemics and that sentinel reporting and model- ademic researchers and a state public health agency and based decision-making are critical for managing limited incorporated it into a Web-based decision-support tool for resources (4–6). For this reason, the United States has pandemic preparedness and response. stockpiled mechanical ventilators in strategically located warehouses for use in public health emergencies, such as iligent preparation and effective countermeasures are an influenza pandemic. The Centers for Disease Control Dcritical to mitigating future influenza pandemics. The and Prevention (CDC) manages this Strategic National 1918 influenza pandemic, the most severe in recent history, Stockpile (SNS) and has plans for rapid deployment to resulted in ≈50 million deaths globally, of which nearly states during critical events (7). 675,000 occurred in the United States (1). The 1957 and However, SNS ventilators might not suffice to meet 2009 pandemics were less severe, causing ≈70,000 and demand during a severe public health emergency. In 9,000–18,000 US deaths, respectively (1). The US Depart- 2002, the SNS included ≈4,400 ventilators (8,9), and ment of Health and Human Services (HHS) estimated (2) 4,500 SNS ventilators were added during 2009 and that 865,000 US residents would be hospitalized during a 2010. The American Association for Respiratory Care moderate pandemic (as in 1957 and 1968) and 9.9 million suggested the SNS inventory should increase to at least during a severe pandemic (as in 1918). 11,000–16,000 ventilators in preparation for a severe in- When severe influenza outbreaks cause high rates of fluenza pandemic10 ( ). The American Association for hospitalization, a surge of medical resources is required, Respiratory Care and CDC (11) provide training on 3 including critical care supplies, antiviral medications, and types of SNS ventilators—LP10 (Covidien, Boulder, CO, USA); LTV1200 (CareFusion, Yorba Linda, CA, Author affiliations: The University of Texas at Austin, Austin, USA); and Uni-vent Eagle 754 (Impact Instrumentation, Texas, USA (H.-C. Huang, G.P. Johnson, P. Damien, Inc., West Caldwell, NJ, USA)—to ensure proper use L.A. Meyers); University of Nebraska, Lincoln, Nebraska, USA nationwide. In addition to the nationally held SNS, some (O.M. Araz); University of Nebraska Medical Center, Omaha, US states maintain their own stockpiles. Nebraska, USA (O.M. Araz); Northwestern University, Evanston, Successful deployment of central ventilator stock- Illinois, USA (D.P. Morton); Department of State Health Services, piles, whether federal or state, requires rapid distribution Austin (B. Clements); Santa Fe Institute, Santa Fe, New Mexico, to healthcare facilities with patients in need, along with ad- USA (L.A. Meyers) equate bed space, requisite supplies, and trained personnel

DOI: http://dx.doi.org/10.3201/eid2306.161417 1Current affiliation: Precima, LoyaltyOne US, Inc., Chicago, IL, USA.

914 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Stockpiling Ventilators for Influenza Pandemics

(12–14). Robust methods for sizing and locating ventilator Methods stockpiles have not yet been developed (15). Wilgis (16) Our approach had 3 stages (Figure 1). First, we estimated the discussed the relative merits of central stockpiling of ven- weekly influenza-related hospitalizations at each site using tilators to be distributed during an emergency versus dis- an adaptive time-dynamic forecasting model. Second, we es- tributing ventilators to hospitals a priori. Centralized stock- timated the number of patients requiring ventilation at each piles benefit from better inventory tracking, more timely site during the peak week on the basis of published estimates repairs, and superior allocation of a limited resource, but of the proportion of hospitalized influenza patients requiring hospital-based supplies facilitate staff training, enable im- mechanical ventilation. Finally, we allocated ventilators at mediate use, and avoid the cost and logistical challenges of minimum cost to achieve a specified level of preparedness central storage and deployment. through a mathematical optimization model. That model as- We developed an optimization framework for allocat- sumed centrally stockpiled ventilators have slightly lower ing mechanical ventilators to central and local stockpiles to treatment rates than locally held ventilators. In the Texas ensure adequate surge capacity during a future pandemic. case study, we estimated hospitalizations under a mild sce- This data-driven method considers the trade-off between nario by fitting the forecasting model to data from the 2009 risk and stockpiling cost, where risk is measured 2 ways: influenza A(H1N1) pandemic, and then we scaled the esti- expected value of unmet demand (EUD; number of influ- mates to simulate moderate and severe pandemics. We sum- enza patients not receiving required ventilation) and proba- marize our optimization model and forecasting methods and bility of unmet demand (PUD; probability at least 1 patient provide details in the online Technical Appendix. does not receive required ventilation). For a given set of healthcare providers in a region, we determined the optimal Optimization Model for Ventilator Stockpiling number of mechanical ventilators to stockpile centrally and Using a 2-stage model, we optimized the allocation of ven- at each provider site. tilators to a central stockpile and several local stockpiles (at As a case study, we considered the US state of Texas healthcare facilities) to ensure that all sites had sufficient under mild, moderate, and severe influenza pandemic sce- surge capacity to manage the peak of an influenza pandem- narios. Based on the Texas Department of State Health ic. We considered the trade-off between unmet ventilator Services (DSHS) response to the 2009 influenza A(H1N1) demand (risk) and the cost of stockpiling ventilators (as- pandemic and planning efforts for future pandemics, we suming cost is proportional to number of ventilators) and considered stockpiling across 9 sites: a centrally held state minimized cost while limiting risk to a specified threshold. stockpile and local stockpiles in each of Texas’ 8 health We analyzed the risk–cost trade-off by solving a family of service regions (HSRs; online Technical Appendix Fig- optimization models, across a range of risk thresholds. ure 1, https://wwwnc.cdc.gov/EID/article/23/6/16-1417- We assumed the following: each stockpiled ventila- Techapp1.pdf). We implemented this model in a Web- tor is both child- and adult-capable, will be used to treat at based decision-support tool for DSHS (17). most 1 patient during peak demand, and will not be used for

Figure 1. Overview of methods for projecting the need to stockpile ventilators for an influenza pandemic, Texas, USA. First, a forecasting model was used to estimate weekly hospitalizations at each site on the basis of historical ILI hospitalization data and CDC ILINet reports. Second, 3 additional factors, along with a spatial correlation coefficient, were used to form a probability distribution for peak-week ventilator demand at each site. Third, an optimization model was solved to determine local and central stockpile allocations and generate trade-off curves between the expected unmet demand and total stockpile and between the probability of unmet demand and total stockpile. CDC, Centers for Disease Control and Prevention; HSR, health service region; ICU, intensive care unit; ILI, influenza-like illness.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 915 RESEARCH noninfluenza patients; stockpiles were established before the Ninth Revision, codes 487 and 488, corresponding to influ- pandemic, and centrally held ventilators can be deployed only enza-like-illness (ILI), and aggregated by HSR. These data once to a site with excess demand (i.e., no redeployment is comprised all Texas hospitals except those in counties with allowed, even though influenza peaks might be asynchronous populations <35,000, those with <100 hospital beds, and across sites); patients requiring ventilatory support cannot those that do not accept insurance or government reimburse- move between sites; locally held ventilators are immediately ment. The number of ILI-related hospital discharges during and successfully administered to on-site patients requiring the 2009 pandemic (April–December 2009) totaled 29,459. care, and centrally held ventilators incur wastage (i.e., a re- We assessed the validity of this International Classification of duced fraction are successfully deployed to healthcare sites Diseases, Ninth Revision–based filter for influenza through upon demand); patients at all sites have equal priority; and comparison with CDC (18) and Texas DSHS (19) reports. consumable ventilator supplies, requisite staffing, and space We also analyzed data from the CDC ILINet, which are in sufficient supply. The optimization model considers tracks weekly outpatient visits related to ILI. CDC guide- expected unmet demand, and we calculated the probability lines define ILI as fever of at least 100°F and cough and/ of unmet demand post hoc, as a secondary risk measure. or sore throat in the absence of a known cause other than influenza. A network of 2,400 sites (health departments, Texas Case Study laboratories, vital statistics offices, healthcare providers, We assumed that ventilators can be stockpiled centrally by and emergency departments) in the 50 states reports to ILI- the Texas DSHS or locally by hospitals in Texas’ 8 HSRs Net, and we obtained weekly reports during the 2009 H1N1 (online Technical Appendix Figure 1). We further assumed pandemic for Texas, aggregated by HSR. Finally, Texas that local stockpiles within an HSR are available throughout DSHS provided data on the 3,730 ventilators stockpiled in the HSR by movement of either ventilators or patients among Texas in 2009 (online Technical Appendix Table 1). healthcare facilities; that is, any patient within an HSR requiring ventilatory support has access to available ventilators Region-to-Region Correlation in Peak Hospitalizations within that HSR. To model peak ventilator demand across For each of the 6 influenza seasons in years starting 2003– Texas’ 8 HSRs under different pandemic scenarios, we 1) 2008 and the 2009 pandemic, we calculated peak-week ILI estimated the region-to-region (HSR-to-HSR) correlation hospitalizations requiring ventilation in each HSR. Across all in peak-week ventilator demand on the basis of 2003–2008 28 pairs of HSRs, the average correlation in peak ventilator seasonal influenza hospitalization data and 2009 pandemic demand was 0.72 ± SD 0.23 (range 0.22–0.98). One HSR, hospitalization data; 2) generated probabilistic estimates of with <3% of total hospitalizations during 2009, had pairwise peak-week influenza-related hospitalizations by fitting our correlations as low as 0.22, but all other pairs of HSRs had forecasting model to a baseline (mild) pandemic scenario coefficients >0.50. We found similar spatiotemporal correla- estimated from 2009 pandemic data; 3) used the estimates tions in hospitalizations when we estimated pairwise HSR- derived in steps 1 and 2 to estimate the numbers of influenza to-HSR correlations for various values of the proportion of patients requiring mechanical ventilation at the pandemic ventilated patients requiring 2 weeks (rather than 1 week) of peak in each HSR; and 4) generated moderate and severe ventilation, and weekly numbers of ILI hospitalizations re- pandemic scenarios by scaling the peak demand estimates quiring ventilation, throughout the 2003–2008 influenza sea- of the mild scenario. We summarize the parameters we used sons and the 2009 pandemic. Given this consistent statewide to estimate peak ventilator demand under different pandemic synchrony in epidemic intensity, we made the simplifying scenarios (Table 1) and outline the data and methods used to assumption that peak hospitalizations in all HSRs were cor- estimate these parameters. related at a pairwise level of 0.70.

Texas Influenza Data Forecasting Model for Hospitalizations We obtained weekly Texas hospital discharge data for 2003– We used a dynamic linear forecasting model (online Tech- 2009, filtered for International Classification of Diseases, nical Appendix), which provides a powerful method for

Table 1. Parameters for estimating peak-week ventilator demand in mild, moderate, and severe influenza pandemics, Texas, USA* Mild (2009- Moderate (1957- and Severe Parameter like) 1968-like) (1918-like) Source Hospitalization scaling over mild 1 3.14 36 (2,21) Proportion of hospitalized ILI patients requiring ICU care 0.2 0.25 0.25 (2,19,21,22,23) Proportion of ICU patients requiring ventilation 0.5 0.5 0.5 (2,21,22) Proportion of ventilated patients requiring 2 weeks of 0.4 0.4 0.4 (22) ventilation Region-to-region correlation for peak-week demand 0.7 0.7 0.7 Estimated *ICU, intensive care unit; ILI, influenza-like illness.

916 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Stockpiling Ventilators for Influenza Pandemics capturing system uncertainty when numerous dynamic that 50% of patients in the ICU who have pandemic influ- factors influence a system 20( ). Although hospitalizations enza require ventilation across all scenarios. could be forecast only on the basis of historical ILI data, our approach can incorporate additional predictors, such as Proportion of Ventilated Patients Requiring 2 Weeks the most recent ILINet reports, to better represent demand of Ventilation uncertainty. Our forecasting method estimated weekly FluSurge 2.0 (22) assumes that ventilatory support of ILI influenza-related hospitalizations in the 8 HSRs for 2009 patients lasts 10 days. We have weekly time resolution pandemic–like scenarios, using CDC ILINet influenza and assumed 60% of patients receiving ventilatory sup- A(H1N1)pdm09 weekly reports as a predictor, from the port require only 1 week, and the remaining 40% require week ending April 4, 2009, through the week ending De- a second week. cember 26, 2009. To account for seasonality, we assumed 5 distinct time periods (September–October, November– Simulating Pandemic Scenarios December, January–February, March–April, and May–Au- We generated a mild scenario by fitting our forecasting gust). We also considered other candidate variables, such model to hospital discharge data for the 2009 pandemic. as school calendars, humidity, and Google Flu Trends, but Because comparable data are not available from 1957 and these did not substantially improve peak estimates. 1968 (moderate) and 1918 (severe), we scaled the 2009 estimates to model these scenarios. HHS (2) and HSC (21) Estimating Regional Ventilator Demand use similar pandemic scaling factors, except HSC rates To estimate regional ventilator peak-week demand, we for hospitalization, ICU care, and mechanical ventilation integrated our weekly forecasts of influenza hospitaliza- are ≈17% and 14% lower than HHS rates for moderate tions in each region, the spatial correlation in peak-week and severe scenarios, respectively. (See [24] for scaling demand for ventilators, and 3 additional factors: 1) the pro- methods for an emerging pandemic.) CDC’s median es- portion of hospitalized ILI patients requiring ICU care, 2) timate of hospitalizations for influenza A(H1N1)pdm09 the proportion of ICU patients requiring ventilation, and 3) (April 2009–April 2010) is 275,000. Combining this with the proportion of ventilated patients requiring 2 weeks of the HHS scenario (online Technical Appendix Table 3), ventilation (rather than 1). To model “spillover” demand we scaled our mild pandemic hospitalization estimates by of patients requiring 2 weeks of ventilation, we used week- 865,000/275,000 = 3.14 and 9,900,000/275,000 = 36 to to-week correlations in influenza hospitalization (online model moderate and severe scenarios, while preserving the Technical Appendix Table 2). variability, spatial correlation, and temporal correlation estimated for 2009. Proportion of Hospitalized ILI Patients Requiring ICU Care From 2009 influenza hospital discharge data, we esti- Results mated that 18% of patients required ICU care during the Under the mild pandemic scenario, recommended stock- peak week. Texas DSHS reported that 23% of the 2,030 piles ranged from 200 to 400 ventilators (Figure 2, panel confirmed influenza A(H1N1)pdm09 patients requiring A). For example, if we specify the risk tolerance to be an hospitalization in Texas during October–December 2009 EUD of at most 5 patients, then the recommended stock- required ICU care (19). For moderate and severe planning pile is 272 ventilators, including a central stockpile of 12. scenarios, the US Homeland Security Council (HSC) (21) The PUD for this scenario, which is computed post hoc, uses an ICU proportion of 15% for the overall pandem- is 30% (Figure 2, panel B). Thus, if the public health de- ic and 25.7% for the peak week. For seasonal influenza, partment builds the recommended central and local stock- CDC’s FluSurge 2.0 (22,23) assumes that a baseline of piles, it can expect that no more than 5 patients statewide 15% of admitted influenza patients require ICU care; HHS will go without ventilation, and a 70% chance exists that makes similar assumptions (2) (online Technical Appendix no demand anywhere will be unmet. As the risk tolerance Table 3). On the basis of these data and reports, we as- decreases from an EUD of 5, the recommended stockpile sumed peak-week ICU proportions of 20% during a mild grows sharply; as the EUD increases, the stockpile de- pandemic and 25% during moderate and severe pandemics. creases nearly linearly (Figure 2, panel A). Ventilators are allocated primarily to local sites rather than to the central Proportion of ICU Patients Requiring Ventilation stockpile (Figure 2, panel C). FluSurge 2.0 assumes 50% of patients with seasonal in- The optimal stockpile allocations under moderate fluenza admitted to the ICU require ventilation 22( ). HSC and severe pandemic scenarios are qualitatively, but not assumes 50% throughout a pandemic (21), and HHS uses quantitatively, similar (Figure 3). With an EUD tolerance 50.4% for a moderate scenario and 50% for a severe scenar- of 5 patients, the recommended stockpiles increase to 1,172 io (2) (online Technical Appendix Table 3). We assumed and 15,697 ventilators for moderate and severe scenarios,

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 917 RESEARCH

Figure 2. Optimal ventilator stockpiles for a mild pandemic scenario, Texas, USA. The total size of the optimal stockpile, summed across the central and 8 HSR stockpiles, decreases as risk tolerance increases. Risk for unmet demand for ventilators is quantified as the expected number of hospitalized influenza patients statewide not receiving necessary ventilation (EUD) (A) and the probability of at least 1 hospitalized patient in Texas not receiving necessary ventilation (PUD) (B). We optimized directly for EUD and calculated PUD post hoc. Red circles indicate EUD/PUD of 5 patients. C) Optimal allocation among central and regional sites when EUD is set to 5 patients, equivalent to a stockpile of 272 ventilators. EUD, expected unmet demand; PUD, probability of unmet demand; HSR, health service region. respectively. These stockpiles scale roughly according to ventilators distributed from the central stockpile is not our assumptions that moderate and severe pandemics have used effectively. This wastage contributes to relatively hospitalization rates of 3.14 and 36 times higher than the small recommended central stockpiles (e.g., just 4.4% of mild pandemic, respectively, and that the fraction of hos- the total stockpile under the mild scenario with an EUD pitalized patients requiring ICU admission increases from of 5 ventilators). As the wastage rate decreases, the cen- 20% in the mild scenario to 25% in the other scenarios. tral allocation slowly increases (Table 2; online Technical Specifically, peak ventilator demand increases by factors Appendix Figure 2). The benefit of risk pooling through a of (0.25/0.20) × 3.14 = 3.93 and (0.25/0.20) × 36 = 45 central stockpile also grows as the region-to-region corre- from the mild to moderate and severe scenarios, respec- lation in peak demand shrinks (Table 2; online Technical tively. This scaling would exactly predict how stockpiles Appendix Figure 3). would grow if we increased the risk tolerance by factors of 3.93 and 45. However, we fixed the EUD limit to 5 Retrospective Analysis of 2009 Pandemic patients, so stockpile growth exceeds these scaling factors. During the 2009 pandemic, hospitals across Texas held an estimated 3,730 ventilators. When aggregated by region, the Sensitivity Analysis 8 HSRs had stockpiles ranging from 151 to 1,233 ventila- We assessed the sensitivity of the recommended stockpil- tors (online Technical Appendix Table 1). Under mild and ing strategies to several factors. For a fixed risk tolerance moderate pandemic scenarios, we projected expected state- (EUD), increasing the proportions of hospitalized patients wide demands for 230 and 903 ventilators, respectively, with requiring ICU admission and ventilation results in compa- each HSR holding a stockpile at least 6 SD above the fore- rable increases in the recommended stockpiles. However, casted mean demand. Given this ample regional surge capac- increasing the proportion of patients requiring 2 weeks of ity, there would have been no need for central stockpiling. ventilation (rather than just 1) produces a slightly more Under the severe scenario, however, the projected statewide complicated effect. Because the demand at peak week will demand is 10,333 ventilators, far exceeding 2009 stockpiles. depend on both established and newly admitted patients, increasing the 2-week proportion from 0 to 1 might not exactly Discussion double the demand. Based on 2009 pandemic hospitalization Central stockpiles can save costs but are advisable only data, peak-week mean demand across Texas is expected to when spatial correlation in peak demand is sufficiently increase by a factor of 1.42 when the 2-week proportion in- low and stockpile deployment is sufficiently reliable. Data creases from 0.4 to 1. The recommended stockpile grows ac- from Texas suggest that influenza peaks strongly correlate cordingly. Under the mild pandemic scenario, the stockpile across regions. Such synchrony undercuts the risk-pooling grows by a factor of 1.38 for an EUD near 0 ventilators and benefits of central stockpiles. Furthermore, successful de- 1.42 for an EUD close to 5 ventilators. ployment requires not only central maintenance and physi- We also varied the wastage rate for centrally held cal transportation of ventilators to patients in need, but ventilators and the region-to-region correlation in peak also healthcare facilities and clinicians trained to admin- demand. The baseline wastage of 0.2 means that 1 in 5 ister and troubleshoot available ventilator models, which

918 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Stockpiling Ventilators for Influenza Pandemics

Figure 3. Optimal ventilator stockpiles for moderate and severe pandemic scenarios, Texas, USA. The total size of the required stockpile, summed across the central and 8 HSR stockpiles, decreases as risk tolerance (EUD) increases, for both moderate (A) and severe (C) pandemic scenarios. For an EUD of 5 patients (red circles), total stockpiles would be 1,172 (A) and 15,697 (C); optimal allocations to central and regional stockpiles are shown for moderate (B) and severe (D) scenarios. EUD, expected unmet demand; PUD, probability of unmet demand; HSR, health service region.

might differ from those held locally. Pandemic-related staff more reliable central stockpile deployment, the more advis- absenteeism might exacerbate this challenge. Our model able a central stockpile. For example, assuming only 0.1% incorporates this limitation by assuming that fraction of wastage, we found that that all ventilators should be held stockpiled ventilators are wasted. When we considered a centrally, regardless of spatiotemporal correlations in peak plausible wastage parameter of 20% (based on discussions demand (Table 2). Thus, as deployment and local capaci- with Texas DSHS about likely impediments to successful ties continue to improve, distance will become less of an deployment), the model recommended that <10% of venti- issue, and the advantages of central stockpiles might out- lators be held centrally. weigh their shortcomings. The recommended allocations among central and local Our surprisingly small central allocation stems from stockpiles hinge critically on the relative efficiencies of a 2 additional factors. First, the uncertainty in our estimates local versus central stockpile, which are largely unknown of peak hospitalizations, based on 2009 pandemic data, is and perhaps changing to favor central stockpiles as deliv- relatively low. Across Texas’ 8 HSRs, the coefficient of ery technology continues to improve. We made the simpli- variation (measuring the level of uncertainty) in peak de- fying assumption that locally held ventilators are perfectly mand for ventilators ranged from 0.17 to 0.36 and averaged matched to patients, and we considered a range of potential 0.24 (online Technical Appendix Table 4). When we in- wastage rates for centrally held ventilators. In general, the crease these coefficients governing uncertainty 3-fold, the

Table 2. Central stockpile size, as a percentage of total stockpile, as a function of wastage and region-to-region correlation in peak ventilator demand during an influenza pandemic, Texas, USA* Central stockpile, % Wastage, % Region-to-region correlation = 0.55 Region-to-region correlation = 0.70 Region-to-region correlation = 0.85 40 2.8 1.5 0.2 30 4.7 2.9 0.9 20 7.0 4.4 2.0 10 9.8 6.7 3.5 1 18.0 13.5 10.1 0.5 25.4 20.6 17.0 0.3 32.8 27.4 22.7 0.2 48.2 43.9 39.3 0.1 100 100 100 *The results in this table are based on a mild pandemic scenario and a limit on expected unmet demand of 5 patients. Bold indicates the baseline value.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 919

RESEARCH recommended central stockpile increases only from 4.4% capacity mismatch due to novel pH1N1 in Canada. Pediatr to 10% of the total, assuming a mild pandemic and a risk Crit Care Med. 2011;12:e51–7. http://dx.doi.org/10.1097/ PCC.0b013e3181e2a4fe tolerance (EUD) of 5 untreated patients. Second, the small 6. Ercole A, Taylor BL, Rhodes A, Menon DK. Modelling the impact central allocation depends on the risk tolerance. As the risk of an influenza A/H1N1 pandemic on critical care demand from tolerance shrinks from an EUD of 5 patients, both the num- early pathogenicity data: the case for sentinel reporting. ber of ventilators in the total stockpile and the percentage Anaesthesia. 2009;64:937–41. http://dx.doi.org/10.1111/ j.1365-2044.2009.06070.x held centrally grow (online Technical Appendix Figure 3). 7. Centers for Disease Control and Prevention. Strategic National Still, even at tighter risk tolerances and a smaller region- Stockpile. 2015 [cited 2016 Jun 16]. http://www.cdc.gov/phpr/ to-region correlation in peak demand of 0.55, the central stockpile/stockpile.htm stockpile is <10%. 8. Malatino EM. Strategic National Stockpile: overview and ventilator assets. Respir Care. 2008;53:91–5, discussion 95. Our retrospective analysis of the 2009 influenza 9. Jamieson DB, Biddison ELD. Disaster planning for the intensive A(H1N1) pandemic in Texas suggests that hospitals had care unit: a critical framework. In: Scales DC, Rubenfeld GD, enough ventilators on hand to treat all patients requiring editors. The organization of critical care. New York: Springer; mechanical ventilation throughout the pandemic. Although 2014. p. 261–75. 10. American Association for Respiratory Care. Guidelines for these quantities are expected to suffice for a moderate acquisition of ventilators to meet demands for pandemic flu (1957- and 1968-like) pandemic, in which hospitalization and mass casualty incidents. 2008 [cited 2016 Jun 16]. rates roughly triple, they would fall far short in a severe https://c.aarc.org/resources/vent_guidelines_08.pdf (1918-like) pandemic. If we optimistically assume perfect 11. American Association for Respiratory Care. The Strategic National Stockpile ventilator training program. 2016 [cited 2016 deployment, that is, 0 wastage, by assuming timely de- Jun 16]. https://www.aarc.org/resources/clinical-resources/ livery, adequately trained and available staff (respiratory strategic-national-stockpile-ventilator-training-program/ therapists, nurses, and physicians), sufficient space to care 12. Meltzer MI, Patel A, Ajao A, Nystrom SV, Koonin LM. Estimates for a potentially large number of patients, and requisite an- of the demand for mechanical ventilation in the United States during an influenza pandemic. Clin Infect Dis. 2015;60(Suppl cillary equipment and supplies, then even a central stock- 1):S52–7. http://dx.doi.org/10.1093/cid/civ089 pile of 8,900 ventilators in Texas—the total number of SNS 13. Ajao A, Nystrom SV, Koonin LM, Patel A, Howell DR, ventilators in 2010 (9)—would fall short, with an expected Baccam P, et al. Assessing the capacity of the US health care unmet demand of 576 patients. system to use additional mechanical ventilators during a large-scale public health emergency. Disaster Med Public Health This study was supported by the National Institutes of Health Prep. 2015;9:634–41. http://dx.doi.org/10.1017/dmp.2015.105 14. Zaza S, Koonin LM, Ajao A, Nystrom SV, Branson R, (Models of Infectious Disease Agent Study grant U01 GM087719- Patel A, et al. A conceptual framework for allocation of 01) and CDC (Public Health Emergency Preparedness). federally stockpiled ventilators during large-scale public health emergencies. Health Secur. 2016;14:1–6. http://dx.doi.org/10.1089/ Dr. Huang is a senior data scientist at Precima, LoyaltyOne US, hs.2015.0043 Inc., Chicago, Illinois, USA. He conducted this research while 15. Timbie JW, Ringel JS, Fox DS, Pillemer F, Waxman DA, a PhD student in the Graduate Program in Operations Research Moore M, et al. Systematic review of strategies to manage and allocate scarce resources during mass casualty events. Ann and Industrial Engineering at the University of Texas at Austin. Emerg Med. 2013;61:677–689.e101. http://dx.doi.org/10.1016/ His research interests include optimization, public health, and j.annemergmed.2013.02.005 supply chain management. 16. Wilgis J. Strategies for providing mechanical ventilation in a mass casualty incident: distribution versus stockpiling. Respir Care. 2008;53:96–100, discussion 100–3. References 17. Texas Department of State Health Services and The University of 1. US Department of Health and Human Services. Pandemic flu Texas at Austin. Texas pandemic flu toolkit. 2013 [cited 2016 Jun history. 2016 [cited 2016 Jun 16]. http://www.flu.gov/pandemic/ 16]. http://flu.tacc.utexas.edu history/index.html 18. Centers for Disease Control and Prevention. CDC estimates of 2. US Department of Health and Human Services. HHS pandemic 2009 H1N1 influenza cases, hospitalizations and deaths in the influenza plan. 2005 [cited 2016 Jun 16]. http://www.flu.gov/ United States, April 2009 through January 16, 2010 by age group. planning-preparedness/federal/hhspandemicinfluenzaplan.pdf 2010 [cited 2016 Jun 16]. http://www.cdc.gov/h1n1flu/estimates/ 3. Sutton J, Tierney K. Disaster preparedness: concepts, guidance, April_January_16.htm and research. Fritz Institute Assessing Disaster Preparedness 19. Texas Department of State Health Services. Texas aggregate Conference; 2006 Nov 3–4; Sebastopol, CA, USA; 2006 [cited surveillance summary—novel influenza A H1N1, week ending 2016 Jun 16]. http://www.fritzinstitute.org/pdfs/whitepaper/ 12/26/09. 2009 [cited 2016 Jun 16]. http://www.dshs.state.tx.us/ disasterpreparedness-concepts.pdf txflu/TX-cumulative-age-archive.shtm 4. Smetanin P, Stiff D, KumarA, Kobak P, Zarychanski R, 20. Petris G, Petrone S, Campagnoli P. Dynamic linear models with Simonsen N, et al. Potential intensive care unit ventilator R. New York: Springer; 2009. demand/capacity mismatch due to novel swine-origin H1N1 in 21. US Homeland Security Council. National planning scenarios version Canada. Can J Infect Dis Med Microbiol. 2009;20:e115–23. 21.3 final draft. 2006 [cited 2016 Jun 16]. http://publicintelligence. http://dx.doi.org/10.1155/2009/808209 net/national-planning-scenarios-version-21-3-2006-final-draft 5. Stiff D, Kumar A, Kissoon N, Fowler R, Jouvet ,P Skip- 22. Zhang X, Meltzer MI, Wortley PM. FluSurge 2.0: a manual to pen P, et al. Potential pediatric intensive care unit demand/ assist state and local public health officials and hospital

920 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Stockpiling Ventilators for Influenza Pandemics

administrators in estimating the impact of an influenza pandemic 24. Reed C, Biggerstaff M, Finelli L, Koonin LM, Beauvais D, on hospital surge capacity (beta test version). 2005 [cited 2016 Jun Uzicanin A, et al. Novel framework for assessing epidemiologic 16]. https://www.cdc.gov/flu/pandemic-resources/tools/downloads/ effects of influenza epidemics and pandemics. Emerg Infect Dis. flusurge2.0_manual_060705.pdf 2013;19:85–91. http://dx.doi.org/10.3201/eid1901.120124 23. Zhang X, Meltzer MI, Wortley PM. FluSurge—a tool to estimate demand for hospital services during the next pandemic influenza. Address for correspondence: Hsin-Chan Huang, Operations Research Med Decis Making. 2006;26:617–23. http://dx.doi.org/ and Industrial Engineering, The University of Texas at Austin, Austin, 10.1177/0272989X06295359 TX 78712, USA; email: [email protected]

February 2015: Complicated Datasets • Entry Screening for • Murine Typhus, Reunion, • Novel Reassortant • Close Relationship Infectious Diseases in France, 2011–2013 Influenza A(H5N8) between Humans Viruses among Inoculated West Nile Virus from • Evidence for Domestic and Wild Ducks, Turkey and Lineage • Timing of Influenza Elizabethkingia anophelis South Korea, 2014 1 Strain from Central A(H5N1) in Poultry and Transmission from Mother African Republic Humans and Seasonal to Infant, Hong Kong • Vesicular Stomatitis Virus– Influenza Activity Based Vaccines against • Ascariasis in Humans Worldwide, 2004–2013 • Microbiota that Affect Risk Lassa and Ebola Viruses and Pigs on Small-Scale for Shigellosis in Children Farms, Maine, USA, in Low-Income Countries 2010–2013

• pH Level as a Marker for • Potentially Novel Ehrlichia Predicting Death among Species in Horses, Patients with Vibrio Nicaragua vulnificus Infection, South Korea, 2000–2011 • Neisseria meningitidis ST-11 Clonal Complex, • Refining Historical Chile 2012 Limits Method to Improve Disease Cluster • Molecular Diagnosis of Detection, New York City, Cause of Anisakiasis in New York, USA • Use of Insecticide-Treated Humans, South Korea House Screens to Reduce • Naturally Acquired Infestations of Dengue • Streptococcus suis • Quantifying Reporting Antibodies against Virus Vectors, Mexico Infection in Hospitalized Timeliness to Improve Haemophilus influenzae Patients, Nakhon Outbreak Control Type a in Aboriginal • Comparative Analysis Phanom Province, Adults, Canada of African Swine Fever Thailand • Tickborne Relapsing Virus Genotypes and Fever, Bitterroot Valley, • Infectious Causes Serogroups • Exposure-Based Montana, USA of Encephalitis and Screening for Nipah Virus Meningoencephalitis in • Awareness and Support Encephalitis, Bangladesh • Simulation Study of the Thailand, 2003–2005 of Release of Genetically Effect of Influenza and Modified “Sterile” Influenza Vaccination on Mosquitoes, Key West, Risk of Acquiring Florida, USA Guillain-Barré Syndrome • Novel Candidatus • Lagenidium giganteum Rickettsia Species Pathogenicity in Detected in Nostril Tick Mammals from Human, Gabon, 2014 • Optimizing Distribution of Pandemic Influenza • Outbreak of Henipavirus Antiviral Drugs Infection, Philippines, 2014 https://wwwnc.cdc.gov/eid/content/ 21/2/contents.htm

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 921 RESEARCH

Invasive Serotype 35B Pneumococci Including an Expanding Serotype Switch Lineage, United States, 2015–2016 Sopio Chochua, Benjamin J. Metcalf, Zhongya Li, Hollis Walker, Theresa Tran, Lesley McGee, Bernard Beall

We used whole-genome sequencing to characterize 199 rare causes of IPD but were geographically widespread in nonvaccine serotype 35B pneumococcal strains that the United States before and after PCV7 introduction (3). caused invasive pneumococcal disease (IPD) in the Unit- One 35B lineage was antimicrobial-susceptible and mul- ed States during 2015–2016 and related these findings to tilocus sequence type (MLST) 452 (35B/ST452), and the previous serotype 35B IPD data obtained by Active Bacte- second strain was penicillin-nonsusceptible 35B/ST558. rial Core surveillance. Penicillin-nonsusceptible 35B IPD During 1995–2001, the penicillin-nonsusceptible 35B/ increased during post–pneumococcal 7-valent conjugate µ vaccine years (2001–2009) and increased further after ST558 lineage, which had resistant MICs of 0.25–2.0 g/ implementation of pneumococcal 13-valent conjugate mL, accounted for 69% of serotype 35B ABCs isolates vaccine in 2010. This increase was caused primarily by (3). During 1999–2007 in the United States, the propor- the 35B/sequence type (ST) 558 lineage. 35B/ST558 and tion of penicillin-nonsusceptible IPD isolates within sero- vaccine serotype 9V/ST156 lineages were implicated as type 35B increased to 84%; 35B/ST558 accounted for this cps35B donor and recipient, respectively, for a single cap- increase (4). Consistent with this observation was a 9-fold sular switch event that generated emergent 35B/ST156 increase in carriage of 35B/ST558 in young children in progeny in 6 states during 2015–2016. Three additional the Atlanta, Georgia, area (5). capsular switch 35B variants were identified, 2 of which After introduction of the 13-valent conjugate vaccine also involved 35B/ST558 as cps35B donor. Spread of 35B/ (PCV13) in 2010, 35B became the most common serotype ST156 is of concern in view of past global predominance in ABCs, associated with MICs >2 µg/mL for penicillin of pathogenic ST156 vaccine serotype strains. Protection against serotype 35B should be considered in next-gener- and amoxicillin in pediatric isolates (6) and in the adult ation pneumococcal vaccines. population (B. Beall, unpub. data). Consistent with its status as a major cause of IPD in the post-PCV13 era, the 35B/ST558 lineage is currently commonly found in dis- lthough the dramatic protective effect of the pneu- ease and asymptomatic pneumococcal carriage in many Amococcal 7-valent conjugate vaccine (PCV7) against countries (7–11). invasive pneumococcal disease (IPD) persisted a full de- We provide a whole-genome sequence (WGS) pipe- cade after its introduction in the United States in 2000, the line−based resolution and description of current 35B lin- emergence of 19A and other non-PCV7 serotypes reduced eages within current ABCs surveillance (6,12), includ- the overall benefit 1,2 ( ). Before PCV7 implementation, ing an invasive 35B variant of the historically successful we observed only 2 different 35B lineages within Active ST156 lineage. Recently, we identified 2 different 35B iso- Bacterial Core surveillance (ABCs) (3), a population- lates recovered during 2009 and 2012 that each appeared to based, multistate program that assesses the effect of inva- have arisen through a unique capsular switch event involv- sive bacterial infections and is part of the Emerging Infec- ing the same 2 parental strains. This observation was made tions Program network of the Centers for Disease Control on the basis of the penicillin-binding protein (PBP) gene and Prevention (CDC; Atlanta, GA, USA) (http://www. types flanking the 35B biosynthetic locus cps35B( ) in each cdc.gov/abcs/index.html). Both lineages were relatively of the variants (6). Only the 35B/ST156 variant detected during 2012 Author affiliation: Centers for Disease Control and Prevention, has emerged and has been detected within 6 states. Atlanta, Georgia, USA ST156 has a remarkable history of conjugate vac- DOI: http://dx.doi/10.3201/eid2306.170071 cine evasion. Formerly the primary genotype of PCV7

922 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Invasive Pneumococci and Serotype Switch Lineage serotype 9V in the United States during the preconju- periods are shown in Table 2. Total numbers of ABCs gate vaccine era (13), 9V greatly decreased after PCV7 35B isolates recovered during 1999–2015 and catego- implementation (1) and was partially replaced by resis- rized by patient age, penicillin MIC, and IPD incidence tant 19A/ST156 (14). To verify that these isolates origi- are shown in Table 3. nated from a single recombinational serotype switch event involving 35B/ST558 and 9V/ST156 parental WGS and WGS-Based Predictions strains, we analyzed regions flanking the cps35B locus Library construction and sequencing was performed as during 2015–2016, 35B/ST156 progeny and the original described (12). WGS accessions for all 199 serotype 35B strain detected during 2012. isolates from 2015–2016, two previous 35B switch strains from previous years, and relevant strains of other sero- Methods types of ST156 from previous years are provided (online Technical Appendix Table, https://wwwnc.cdc.gov/EID/ Isolates article/23/6/17-0071-Techapp1.pdf). WGS pipeline data The surveillance population of ABCs is ≈32 million per- and quality metrics for all isolates are also provided (on- sons in 10 states (http://www.cdc.gov/pneumococcal/ line Technical Appendix Table). Capsular serotypes, an- surveillance.html). Serotype 35B IPD isolates described timicrobial genotypes/phenotypes, MLST, sequence type include 132 recovered during 2015 and 67 recovered dur- (ST), and pili (presence or absence) for year 2015–2016 ing ABCs in 2016 (Table 1, https://wwwnc.cdc.gov/EID/ isolates were deduced through our bioinformatics pipeline article/23/6/17-0071-T1.htm). The listing of 35B iso- (6,12,15). lates from 2016 is incomplete because we typically receive all ABCs isolates recovered during a given year by Phylogeny the following summer. Relevant ST156 lineage isolates Paired-end fastq files were trimmed with Cutadapt- ver of other serotypes recovered during this and previous sion 1.8.1 (17), and draft genome assemblies were

Table 2. Nonserotype 35B isolates of ST156 lineage included in study of penicillin-nonsusceptible 35B pneumococcal isolates causing IPD, United States, 1998–2015* Serotype/ Antimicrobial resistance phenotype, MIC, g/mL¶ MLST type Non-PBP resistance Cli + State (year (no.)† No. PBP type‡ determinants§ Pen Amo Tax Cft Cfx Mer Ery Tet Cot Fq isolated) 9V/156 (25) 12 15:12:18 folAI100L, 4 2 1 2 >2 0.5 S S R S CA, GA, MD, folPins178 MN, NY, TN (1998–1999) 9 15:12:18 mef, folAI100, 4 2 1 2 >2 0.5 R S R S CA, CT, MD, folPins178 MN, OR, TN (1998,1999, 2015) 2 15:12:18 ermB, tetM, 4 2 1 2 >2 0.5 R R R S CA (2015) folAI100L, folAins178 1 15:12:18 mef, tetM, folAI100L, 4 2 1 2 >2 0.5 R S R S CT (2015) folPins178 1 15:12:228 Mef, folAI100L, 4 2 1 2 >2 0.5 R S R S MD (2016) folPins178 19A/156 (4) 3 29:12:26 mef, folAI100L, 4 2 8 4 >2 0.5 R S R S CA,GA (2009) folAins189 1 8:12:36 mef, folAI100L, 1 2 0.5 0.5 2 0.25 R S R S GA (2015) folAins189 13/156 (1) 1 15:12:173 folAI100L, 1 1 0.25 0.25 1 0.5 S S R S TN (2015) folPins178 31/156 (1) 1 15:12:18 mef, folAI100L, 4 2 1 2 >2 0.5 R S R S MN (2015) folPins178 *All isolates were positive for pilus PI-type 1 and negative for pilus PI-type 2. IPD, invasive pneumococcal disease; MLST, multilocus sequence type; PBP, penicillin-binding protein; R, resistant; S, susceptible; ST, sequence type. †Types that probably arose through serotype switching are indicated in bold. ‡See Li et al. (15) and MIC correlates for PBP types (http://www.cdc.gov/streplab/mic-tables.html). §For a description of WGS-based bioinformatic pipeline for deduction of all features shown, see Metcalf et al. (6,12). For a description of folP insertions (folPins178, folP189), see Figure 1 in Metcalf et al. (12). ¶Predicted MICs for β-lactam antimicrobial drugs were based on transpeptidase domain sequences of PBPs 1a, 2b, and 2x (http://www.cdc.gov/streplab/mic-tables.html). For penicillin (meningitis only), nonsusceptible is considered >0.12 g/mL (16). Currently applied clinical cutoffs are also provided for the other 5 β-lactams shown (16). Where shown, R and S correspond to breakpoint MIC values (16). Amo, amoxicillin; Cft, ceftriaxone; Cfx, cefuroxime; Cli, clindamycin; Cot, cotrimoxazole; Ery, erythromycin; Fq, fluoroquinolones levofloxacin and ciprofloxacin; Mer, meropenem; Pen, penicillin; Tax, cefotaxime.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 923

RESEARCH

Table 3. Annual incidence and proportions of penicillin-nonsusceptible 35B pneumococcal isolates causing IPD, United States, 1999–2015* No. 35B isolates from penNS 35B Surveillance patients by age, y Relative incidence Pen MIC, g/mL isolates/penS 35B Year population % CIS <5 ≥5 of 35B IPD† >2 0.121 <0.06 isolates 1998 17,383,935 86.1 3 18 1.40 9 5 7 2.0 1999 18,550,681 87.0 2 18 1.24 8 3 9 1.2 2000 19,821,607 86.3 4 21 1.46 7 9 9 1.8 2001 22,479,308 88.1 2 40 2.12 14 18 10 3.2 2002 25,051,246 87.6 2 34 1.64 20 8 8 3.5 2003 25,264,246 91.4 11 49 2.56 22 26 11 4.4 2004 27,419,898 87.9 15 69 3.49 44 25 15 4.6 2005 27,816,784 89.5 11 57 2.73 35 18 15 3.5 2006 28,204,455 86.7 1 65 2.70 37 15 14 3.7 2007 28,579,312 87.5 5 83 3.52 51 23 14 5.3 2008 28,856,774 86.7 12 80 3.68 62 16 14 5.6 2009 29,206,528 89.8 8 70 2.97 45 16 17 3.6 2010 29,757,552 90.2 4 67 2.65 52 14 5 13.2 2011 30,075,050 90.3 11 77 3.28 71 7 10 7.8 2012 30,356,544 90.6 13 101 4.14 94 14 6 18.0 2013 30,604,240 88.7 15 114 4.75 94 25 10 11.9 2014 31,328,211 88.2 16 116 4.78 104 22 6 21.0 2015 31,977,800 92.0 10 121 4.45 101 24 6 20.8 *CIS, case isolates serotyped; IPD, invasive pneumococcal disease. †Estimated cases/million = total 35Bs x 100/% CIS surveillance population/1,000,000. constructed by using VelvetOptimiser version 2.2.5 with Population Snapshot of Ongoing ABCs for an optimal kmer value calculated by using VelvetK (18). 35B IPD, 2015–2016 Core genome single-nucleotide polymorphism (SNP) Among 2,710 IPD isolates obtained during 2015 and sub- identification and alignment were performed - byus jected to WGS, 132 (4.9%) were serotype 35B. Of 1,528 ing kSNP3.0 (19). A maximum-likelihood phylogenetic IPD isolates recovered from partial year 2016 IPD surveil- tree was generated from the core SNP alignment by us- lance, 67 (4.4%) were serotype 35B. Most (168/199) of ing RaxML version 7.3.0 (20). RaxML was run with an these isolates belonged to penicillin-nonsusceptible clonal ASC_GTRGAMMA DNA substitution model and used complex (CC) 558 (168 isolates) and CC156 (21 isolates) the Lewis method for ascertainment bias correction. Node (Figure 1; Table 1). Serotype 35B CC558 and CC156 iso- support was assessed by using 500 bootstrap replicates. lates of all serotypes discussed were uniformly positive for the rrgA gene (Tables 1, 2), which encodes a pilus sub- Conventional MIC Testing and Serotyping unit that functions in epithelial adhesion (22). Ten isolates Serotype 35B isolates recovered during 2015 were subject- of long-standing penicillin-susceptible 35B/ST452 (3) ed to conventional broth dilution testing for determination were also recovered. Single 35B isolates were identified of antimicrobial MICs. A selection of these isolates were of ST1092, a lineage of conjugate vaccine serotypes 6A also subjected to conventional serotyping by using CDC and 6B (http://pubmlst.org/spneumoniae/) and of ST11818 typing antisera as described (6). (highly related to 15A/ST63), an antimicrobial-resistant nonvaccine serotype lineage that has increased in the post- Statistical Analyses conjugate vaccine era (4). A χ2 test was performed to evaluate differences among groups. This test was performed by using OpenEpi Version CC558 (35B/CC558) 3.01 (http://www.openepi.com/Menu/OE_Menu.htm). Of 168 35B/CC558 isolates obtained, 147 were ST558, 20 were single-locus variants (SLVs) of ST558 correspond- Results ing to 13 STs, and 1 was a double-locus variant (Figure 1; Table 1). Within CC558, only ST558 had SLVs, which Increase in Penicillin-Nonsusceptible 35B is consistent with initial successful establishment of 35B/ during the Conjugate Vaccine Era ST558 in its ecologic niche and subsequent rare shedding During 1998–2001, penicillin-nonsusceptible 35B ac- of closely related SLVs (21). counted for 67.6% (108) of serotype 35B ABCs isolates The increased incidence of 35B IPD during the post- (Table 3). During 2002–2015, the proportion of penicillin- PCV7 period (2.1–3.7 cases/million population during 2001– nonsusceptible IPD isolates with serotype 35B increased to 2009 vs. 1.2–1.3 cases/million population during 1998– 87.7% (1,237; p<0.001). 1999) and the post-PCV13 period (3.3–4.8 cases/million

924 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Invasive Pneumococci and Serotype Switch Lineage

Figure 1. Population snapshot of 199 serotype 35B pneumococcal isolates obtained by ongoing Active Bacterial Core surveillance, United States, 2015–2016, configured by using eBURST (21). Diameters are proportional to number of isolates. Solid lines indicate single-locus variants, and the single dashed line indicates a double-locus variant of ST558. ST, sequence type.

population during 2011–2015), combined with the consis- folPins178) strain serving as the recipient strain for a 35B/ tent trend of markedly increased proportions of penicillin- ST558 cps35B donor strain (Figure 3). Another potential nonsusceptible 35B IPD isolates throughout the conjugate recipient strain present before and after PCV13 introduc- vaccine era (Table 2), is consistent with reported increased tion was 19A/ST156 (6,14). However, this lineage is as- 35B/ST558 in IPD and carriage (4–11). ABCs surveillance sociated with the folPins189 insertion (Table 2; Figure 2). sites increased after 2000, but 35B IPD incidence calcula- Both flanking pbp loci from 35B/ST558 were cotrans- tions did not vary whether including the expanded surveil- ferred with the cps35B locus to replace the cps9V, pbp1a-15 lance sites or by using only continuously participating ABCs and pbp2x-18 determinants in the putative 9V/ST156 re- sites during 1998–2015. cipient, which resulted in PBP type 4:12:7 (Figure 3). This serotype switch progeny strain was obtained from a 4-year- CC156 (35B/CC156) old child during 2012 and is a potential progenitor of the We analyzed PBP types (6,12,15) of 35B/ST558 (4:7:7) and current invasive 35B/CC156 lineage (isolate 2012221165) 35B/ST156 (4:12:7) isolates. These PBP amino acid sequence (Figures 2, 3). Antimicrobial resistance markers PBP2b-12, types are used for predicting β-lactam MICs and correspond mef, FolA-I100L, and folPins178, combined with the close to PBP transpeptidase domains from PBP1a, PBP2b and PB- phylogenetic relatedness of the 9V/ST156 isolates (Figure P2x, respectively. PBP genes pbp1a and pbp2x flank opposite 2), suggest that a member of this lineage served as the re- ends of the capsular biosynthetic locus and are sometimes co- cipient parental strain for the 35B/ST156 clade isolated in transferred during serotype switching events (6,23–25). The ongoing ABCs during 2015–2016 (Figure 3). 35B/ST558 lineage has been nearly exclusively associated Analysis of the regions flanking the cps35B locus for with PBP type 4:7:7 among isolates obtained since 1998, and all 35B/ST156 lineage isolates obtained during 2015–2016 the serotype 9V/ST156 lineage is similarly highly associated showed identical recombinational sites at bases 6,453 (left with PBP type 15:12:18 (6,15). However, 9V/ST156 is rare coordinate of progeny reference) (Figure 3) and 10,836 among IPD isolates in the post-PCV7 period. (right coordinate), which is clearly indicative of a single In addition to the PBP2b-12 marker, mef gene, and event within a 35B/ST156 ancestral strain of the 19 prog- FolA-I100L substitution, candidate cps35B recipient 9V/ eny shown (Figure 3). Thus, a double-crossover event re- ST156 strains contain the 2-codon insertion designated fol- placed the recipient strain cps9V locus and flanking PBP Pins178 (Table 2; Figures 2, 3). Such 1–2 codon folP inser- markers (2x-18 and 1a-15) with the cps35B locus and its tions, which together with FolA-I100L confer cotrimoxa- flanking PBP markers (2x-7 and 1a-4). On the basis of zole resistance, are categorized by specific location of the available strain data, the original progeny strain is predict- insertion and specific sequence flanking the insertions (12). ed to have been an ST156 strain with the PBP type 4:12:7; a These genomic features are also found within the 35B/ total of 14 of the 19 35B/ST156 lineage strains still shared ST156 lineage isolates described (Table 1; Figures 2, 3), these characteristics (Table 1). Five isolates are SLVs or which are consistent with a 9V/ST156 (mef, FolA-I100L, differ in PBP2b type.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 925 RESEARCH

Figure 2. Phylogenetic analysis of potential recipient and serotype switch pneumococcal progeny strains within the ST56 lineage based upon a total alignment of 10,409 core single-nucleotide polymorphisms, United States, 2015–2016. All 20 serotype 35B progeny shown were recovered through Active Bacterial Core surveillance, and all but 2 indicated strains were obtained during 2015–2016. Isolate features are depicted for the 2 major nodes, with exceptions indicated by asterisks within the tree. Bootstrap values are indicated at key nodes. The serotype 9V isolate that was used for the reference recipient sequences described in Figure 3 is indicated as the third isolate from bottom. All isolates above the 9V recipient reference within this major cluster, except where indicated, are also 9V/ST156. Three additional serotype switch ST156 strain types detected by Active Bacterial Core surveillance during 2015–2016 are indicated by asterisks (single isolates of serotypes 31 and 13 and 4 serotype 19A isolates). Scale bar indicates nucleotide substitutions per site.

We detected the small (491 bp) segment (bases 11349– a highly penicillin-resistant ST320 strain, prevalent dur- 11839 of progeny) (Figure 3) of clearly recipient lineage ing the post-PCV7 era and having PBP type 13:11:16 (6), origin within the left side of the major recombinational simultaneously replaced the pbp2b locus and flanking ddl fragment. During a single–double crossover event that fa- sequence to change the PBP type to 4:11:7 and the MLST cilitates a serotype switch, additional independent double- type to ST12921 through transfer of ddl with the resistance- crossover events appear to occur concurrently (27). How- conferring selectable pbp2b allele (28). We also observed ever, these events probably do not occur simultaneously. an increased MIC for amoxicillin for this PBP type 4:11:7 It appears that the actual serotype switch event involved a strain compared with MICs for PBP type 4:12:7 strains shorter donor fragment bordering upon the right side of this (Table 1). small recipient lineage segment (base 11839), followed by Two isolates (20152877 and 20161763) underwent a second double-crossover event that bordered upon the left a postswitch intra–cps35B gene deletion event within side of the recipient lineage fragment (base 11349). the wciG gene (Figure 3), which is predicted to encode A single penicillin-susceptible 35B/ST162 (SLV an acetyltransferase (26). Although these 2 isolates were of ST156) has the completely sensitive PBP type 0:0:0 serotyped as 35B by using CDC typing antisera, they (6,12) (Table 1). This strain arose through an independent differed in reactivity with serologic factor 35a compared serotype switch event that involved a penicillin-susceptible with the other 17 isolates of this lineage (Table 4). Typ- recipient strain. ing antisera factors 29b and 35c are the CDC Quellung reagents definitive for serotype 35B. The original proto- Postserotype Switch Event Diversification of col (29,30) that CDC first followed for serogroup 35 res- 35B/ST156 Progeny olution also used factor 35a along with factors 29b and Five of the 19 35B/ST156 lineage progeny showed indi- 35c for identification of serotype 35B. We found that cations of genetic diversification that occurred after the the 2 wciG deletion strains did not react with factor 35a, capsular switch event. Four of these isolates have 1 of 3 but the other 17 serotype 35B strains reacted strongly SLV MLSTs of ST156 (ST9910, ST11584, and ST12921) with factor 35a. These preliminary data is suggestive of (Table 1). Although 18 of the 19 strains were PBP type a new serotype within serogroup 35 because this specific 4:12:7, the SLV ST12921 variant had PBP type 4:11:7. For factor reactivity pattern has not been observed for sero- this particular strain, it is probable that recombination with group 35 (31).

926 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Invasive Pneumococci and Serotype Switch Lineage

Figure 3. Diagrammatic representation of cps loci and adjacent regions from donor, recipient and progeny strains depicting serotype switch event for pneumococcal isolates, United States, 2015–2016. Red and green lines in progeny indicate regions of sequence identity or near identity (<2 single-nucleotide polymorphisms/10,000 bp) to the above corresponding donor and recipient sequences, respectively. Rectangles indicate relative locations of PBP gene types for pbp2x and pbp1a. Below each cps locus, a representative reference strain is indicated along with relevant features determined through a bioinformatics pipeline (MLST, PBP type, resistance markers). Junctions between donor and recipient sequences involved in the 2 single recombinational crossovers in the gene replacement event are indicated with blue arrowheads above the progeny diagram, although a single short internal region with sequence identity to the recipient nested within the donor fragment (left coordinates 11349–11839) is also present. Below each green or red segment of the progeny, the level of sequence identity to donor and recipient is provided. The list of each progeny strain, date of isolation, and state is provided. Where MLST is not ST156, its single locus variants (ST9910, ST11584, and ST12921) are included. Two exceptions indicating flanking post-switch recombination within left coordinates 1–6453 are indicated in isolates 20152884 and 20161413 (asterisks): isolate 20152884 had only 99.3%−99.5% identity to recipient and donor over bases 1–3715, and isolate 20161413 had only 99.5%–99.7% identity to recipient and donor over bases 1–2143. Two strains on the right indicate a post-switch deletion event within the wciG putative acetyltransferase gene, which putatively contributes to the acetylation pattern of the serotype 35B polysaccharide (26). MLST, multilocus sequence type; PBP, penicillin-binding protein; ST, sequence type.

Further indication of chromosome-wide postswitch differs in the flankingpbp2x marker and distal pbp2b mark- diversification of this single clade was shown in the left- er (PBP type 4:7:18). Features of this variant have been flanking region of the cps35B locus. Two progeny strains described (6), and we have not obtained additional 35B showed diversification within the first 2–3.7 kb when com- isolates with these distinguishing features. A third sero- pared with the other 17 progeny. The 6-kb region immedi- type switch event involving a CC156 recipient strain is in- ately to the left of base 1 in all of the progeny strains had tuitive from the pipeline data, which indicate that the 35B <99.2% sequence identity with the most similar potential SLV of ST162 is featured by the β-lactam–susceptible ST156 parental recipient strains that we analyzed. How- PBP type 0:0:0 (Table 1). ST162 has long been associated ever, beyond this segment, progeny had sequence identity with penicillin-susceptible 9V strains (13) and more re- with the parental strain for >8 kb. cently with PI-1 positive and penicillin-susceptible 23B, 15B, and 15C strains (6). 35B/ST156 Variant Lineages Arising through Separate Serotype Switch Events Nonserotype 35B Variants of ST156 The 9V/ST156 clade also appears likely to have served Single ST156 isolates of serotypes 31 and 13 were obtained as the recipient for an independent serotype switch from during 2015 (Figure 2) and showed high relatedness to dif- the same 2 parental strains (Figure 3) which resulted in ferent 9V subdivisions. Again, the likely recipient back- 35B/ST10174, an SLV of ST156 (Figure 2) obtained from ground for the presumed capsular switch does not appear an infant during 2009 (isolate 2009219987). This isolate likely to involve the 19A/ST156 lineage (Figure 2), which

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 927 RESEARCH

Table 4. Serologic comparison of 35B/ST156 lineage strains sampling of isolates from all age groups during 1998–2013 with CDC Quellung reagents for resolution of serogroup 35B (4,6,13; B. Beall, unpub. data). invasive serotype pneumococci including an expanding Before 2015, we detected only 1 isolate of the 35B/ serotype switch lineage, United States, 2015–2016* Quellung factor ST156 lineage that was recovered during 2012 (6). Thus, Strain 35b 29b 35c 42a 35a we feel justified in describing it as newly emergent iso- cps35B†  + +  + late. A smaller study was recently published (during peer cps35B (wciG deletion)‡  + +   review of this article) that described 78 invasive and 48 *CDC, Centers for Disease Control and Prevention; ST, sequence type; , negative; +, positive. noninvasive serotype 35B isolates obtained during 1994– †Refers to 17 progeny resulting from recombination indicated in left 2014 from 8 hospitals in 8 states (33). Our data, which column under progeny lineage diagram in Figure 3. ‡Refers to 2 indicated wciG deletion isolates in right column under included a population-based sampling of 199 35B isolates progeny lineage diagram in Figure 3. obtained during 2015–2016, clearly shows the current national predominance of 35B/ST558 and does not support was well-represented during the 2000s after PCV7 imple- the observation that 35B/ST156 is the major contributor mentation (6,14,32). to post-PCV13 antimicrobial-resistant 35B. Both studies noted the initial appearance and emergence of 35B/ST156 CCs of Remaining 35B Isolates Obtained during in the post-PCV13 period. 2015–2016 This recent identification of the antimicrobial-resistant Ten of the 12 isolates other than penicillin-nonsusceptible 35B/ST156 lineage and its subsequent detection within 6 CC558 and CC156 (together composed of 187 isolates) ABCs sites is a cause for concern. The ST156 lineage has were of the long-established penicillin-susceptible 35B/ shown a remarkable propensity to persist through undergo- ST452 lineage (3), which decreased in proportion during ing serotype-switch events (12,23,25,32). The penicillin- the 2000s (4) (Figure 1; Table 3). The 2 remaining 35B resistant 9V/ST156 lineage was the predominant serotype isolates obtained during 2015–2016 also appear to have 9V cause of IPD in the United States during the pre-PCV7 originated through serotype switching events involving a era (13,32). Soon after introduction of PCV7, serotype 9V 35B/ST558 cps35B donor, as implicated by the presence IPD became rare (1,2), and 19A became the predominant of the PBP1a-4 or PBP2x-7 determinants flanking the cps representative of the ST156 lineage within ABCs (14,32). loci of these 2 progeny strains (35B/ST11818 and 35B/ After introduction of PCV13, 35B has become the pre- ST1092) (Table 1). The 35B/ST11818 variant is an SLV dominant serotype of the ST156 lineage within the United of ST63 and has the same resistance features and accessory States (B. Beall, unpub. data). resistance genes (ermB and tetM) as the currently common A distinct antimicrobial-susceptible serotype 35B SLV 15A/ST63 clone (4,6,32). The 35B/ST1092 isolate is likely of ST156 (35B/ST162) is included among 35B ABCs of to have originated through serotype switching with a sero- 35B during 2015–2016 by the β-lactam–susceptible PBP group 6 recipient strain (6). The 14 remaining ST1092 iso- type 0:0:0. Thus, our data indicates that >3 independent lates obtained during surveillance in 2015–2016 were sero- serotype switches involving the nonvaccine type cps35B type 6C. The 6 previously collected ST1092 IPD isolates locus and the broad ST156 clonal complex serving as re- (1999–2013) represented in our WGS collection are from cipient strains have previously occurred. In this study, we serotype 6A, 6B, and 6C strains (6; B. Beall, unpub. data). demonstrated that all penicillin-nonsusceptible 35B/ST156 lineage isolates obtained during current ABCs (2015–2016) Discussion arose through a single ancestral recombination event. This An increase of penicillin-nonsusceptible serotype 35B IPD event was facilitated through detailed analyses of crossover and carriage caused by 35B/ST558 has been apparent in the points and comparisons of corresponding regions of all United States since the introduction of PCV7 in 2000, and progeny isolates with likely parental 35B/ST558 and 9V/ it has shown a major increase after PCV13 implementa- ST156 strains. The genetic plasticity of the ST156 lineage tion (4–8). This finding is of concern because even strains is also highlighted in this study by detection of postsero- that are rarely detected in IPD sometimes rapidly emerge. type switch changes affecting β-lactam resistance (PBP1a For example, the 19A/ST320 strain was not detected dur- type) and capsular serotype (wciZ deletion), which is po- ing extensive characterization of pre-PCV7 ABCs isolates tentially reflective of recent antimicrobial drug pressure (13), yet it became the predominant invasive pneumococ- and immunologic selection pressure. cal strain during 2005–2009 (14,32). We have performed An additional 35B variant within a vaccine serotype comprehensive strain characterization (MLST and WGS) lineage is shown with ST1092 that is typically associated of pediatric (from children <5 years of age) ABCs isolates with serogroup 6 strains (Figure 1). Because these puta- obtained during 1999, 2001, 2002, 2008, 2009, and 2011– tive 35B switch variants were not detected during exten- 2013 (6,13) and WGS-based characterization of a large sive strain surveillance before and shortly after conjugate

928 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Invasive Pneumococci and Serotype Switch Lineage vaccine implementation (3,4,13), it is plausible that these obtained during 1998–1999, 2009, and 2012; the Minnesota serotype switches occurred after implementation of conju- Department of Public Health laboratory for performing gate vaccine. The observation of a 35B variant within the susceptibility testing of all isolates from Minnesota; and antimicrobial 15A/ST63 lineage brings the number of sero- Robert E. Gertz, Jr for providing insights into CDC type switch events generating 35B strains described in this pneumococcal typing serum factors. study to 5; (35B/ST11818, 35B/ST156, 35B/ST162 from This study was supported by the Centers for Disease Control and 2015–2016, and 35B/ST10174 from 2009). Except for the Prevention and used data from the Streptococcus pneumoniae 35B/ST162 variant, these serotype switch events were pre- MLST website (http://pubmlst.org/_spneumoniae) at the dicted on the basis of progeny PBP type to involve the 35B/ University of Oxford (Oxford, UK). ST558 strain as the cps35B donor (online Technical Ap- pendix Table). Development of the Streptococcus pneumoniae MLST website Although conjugate vaccines have a history of provid- was supported by the Wellcome Trust. ing effective and durable protection against IPD 1,2( ), the Dr. Chochua is a researcher in the Streptococcus Laboratory, continued emergence and expansion of serotype 35B into National Center for Immunization and Respiratory Diseases, different clonal complexes supports continued development Centers for Disease Control and Prevention, Atlanta, GA. of wider spectrum pneumococcal vaccines. Serotype 19A Her research interests are next-generation sequencing, IPD, although relatively uncommon in the pre-PCV7 era, characterization of clinical streptococcal isolates, antimicrobial rapidly became the predominant invasive serotype in the resistance, genetic adaptations, and outbreak responses. post-PCV7 period (14,32,34). Serotype 35B strains have several of the same features that were found among serotype 19A strains before implementation of PCV7 in 2000. References These features that could predispose for serotype 35B to 1. Pilishvili T, Lexau C, Farley MM, Hadler J, Harrison LH, Bennett NM, et al.; Active Bacterial Core Surveillance/Emerging continue its increasing trend as a cause of IPD include its Infections Program Network. Sustained reductions in invasive lack of inclusion within conjugate vaccine, high carriage pneumococcal disease in the era of conjugate vaccine. J Infect Dis. rates within children, antimicrobial resistance, clonal ex- 2010;201:32–41. http://dx.doi.org/10.1086/648593 pansion, and serotype switching. An experimental 15-va- 2. Moore MR, Link-Gelles R, SchaffnerW, Lynfield R, Lexau C, Bennett NM, et al. Effect of use of 13-valent pneumococcal lent conjugate vaccine in development includes serotypes conjugate vaccine in children on invasive pneumococcal disease in 22F and 33F (35), which have increased as causes of IPD children and adults in the USA: analysis of multisite, population- in the postconjugate vaccine era. Serotypes 15A, 15B, and based surveillance. Lancet Infect Dis. 2015;15:301–9. 23A are expressed by moderately antimicrobial-resistant http://dx.doi.org/10.1016/S1473-3099(14)71081-3 3. Beall B, McEllistrem MC, Gertz RE Jr, Boxrud DJ, Besser JM, clones and are not uncommon causes of IPD (4,32). Al- Harrison LH, et al.; Active Bacterial Core Surveillance/Emerging though less resistant to β-lactam antimicrobial drugs than Infections Program Network. Emergence of a novel penicillin- 35B/ST558 and 35B/ST156, these strains also present a nonsusceptible, invasive serotype 35B clone of Streptococcus challenge to address through more encompassing pneumo- pneumoniae within the United States. J Infect Dis. 2002;186:118– 22. http://dx.doi.org/10.1086/341072 coccal vaccines. 4. Gertz RE Jr, Li Z, Pimenta FC, Jackson D, Juni BA, Lynfield R, et al. Increased penicillin-nonsusceptibility of nonvaccine Acknowledgments serotype (other than 19A and 6A) invasive pneumococci in We thank A. Reingold, S. Brooks, H. Randel, L. Miller, post 7 valent conjugate vaccine era. J Infect Dis. 2010;201:770–5. http://dx.doi.org/10.1086/650496 B. White, D. Aragon, M. Barnes, J. Sadlowski, S. Petit, 5. Sharma D, Baughman W, Holst A, Thomas S, Jackson D, M. Cartter, C. Marquez, M. Wilson, M. Farley, S. Thomas, da Gloria Carvalho M, et al. Pneumococcal carriage and invasive A. Tunali, W. Baughman, L. Harrison, J. Benton, T. Carter, disease in children before introduction of the 13-valent R. Hollick, K. Holmes, A. Riner, Ruth Lynfield, Anita Glennen, conjugate vaccine: comparison with the era before 7-valent conjugate vaccine. Pediatr Infect Dis J. 2013;32:e45–53. C. Holtzman, R. Danila, K. MacInnes, K. Scherzinger, http://dx.doi.org/10.1097/INF.0b013e3182788fdd K. Angeles, J. Bareta, L. Butler, S. Khanlian, R. Mansmann, 6. Metcalf BJ, Gertz RE Jr, Gladstone RA, Walker H, Sherwood LK, M. Nichols, N. Bennett, S. Zansky, S. Currenti, S. McGuire, Jackson D, et al.; Active Bacterial Core surveillance team. Strain A. Thomas, M. Schmidt, J. Thompson, T. Poissant, W. Schaffner, features and distributions in pneumococci from children with invasive disease before and after 13-valent conjugate vaccine B. Barnes, K. Leib, K. Dyer, L. McKnight, R. Gierke, implementation in the USA. Clin Microbiol Infect. 2016;22:60. O. Almendares, J. Hudson, L. McGlone, Tamara Pilishvili, e9–29. http://dx.doi.org/10.1016/j.cmi.2015.08.027 G. Langley, and Cynthia Whitney for their contributions to 7. Desai AP, Sharma D, Crispell EK, Baughman W, Thomas S, establishment and maintenance of the ABCs system; Rebecca Tunali A, et al. Decline in pneumococcal nasopharyngeal carriage of vaccine serotypes after the introduction of the 13-valent Gladstone, Stephen Bentley, and Paulina Hawkins for providing pneumococcal conjugate vaccine in children in Atlanta, Georgia. genome sequences through the Global Pneumococcal Pediatr Infect Dis J. 2015;34:1168–74. http://dx.doi.org/10.1097/ Sequencing Project (http://www.pneumogen.net) from isolates INF.0000000000000849

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 929 RESEARCH

8. Kaur R, Casey JR, Pichichero ME. Emerging Streptococcus 22. Barocchi MA, Ries J, Zogaj X, Hemsley C, Albiger B, Kanth A, pneumoniae strains colonizing the nasopharynx in children after et al. A pneumococcal pilus influences virulence and host inflam- 13-valent pneumococcal conjugate vaccination in comparison to matory responses. Proc Natl Acad Sci U S A. 2006;103:2857–62. the 7-valent era, 2006–2015. Pediatr Infect Dis J. 2016;35:901–6. http://dx.doi.org/10.1073/pnas.0511017103 http://dx.doi.org/10.1097/INF.0000000000001206 23. CoffeyTJ, Daniels M, Enright MC, Spratt BG. Serotype 14 9. Kawaguchiya M, Urushibara N, Kobayashi N. Multidrug resistance variants of the Spanish penicillin-resistant serotype 9V clone of in non-PCV13 serotypes of Streptococcus pneumoniae in northern Streptococcus pneumoniae arose by large recombinational Japan, 2014. Microb Drug Resist. 2017;23:206–14. http://dx.doi.org/ replacements of the cpsA-pbp1a region. Microbiology. 1999; 10.1089/mdr.2016.0054 145:2023–31. http://dx.doi.org/10.1099/13500872-145-8-2023 10. McElligott M, Vickers I, Meehan M, Cafferkey M, Cunney R, 24. Brueggemann AB, Pai R, Crook DW, Beall B. Vaccine escape Humphreys H. Noninvasive pneumococcal clones associated with recombinants emerge after pneumococcal vaccination in the antimicrobial nonsusceptibility isolated from children in the era of United States. PLoS Pathog. 2007;3:e168. http://dx.doi.org/ conjugate vaccines. Antimicrob Agents Chemother. 2015;59:5761– 10.1371/journal.ppat.0030168 7. http://dx.doi.org/10.1128/AAC.00990-15 25. Wyres KL, Lambertsen LM, Croucher NJ, McGee L, 11. Golden AR, Adam HJ, Gilmour MW, Baxter MR, Martin I, von Gottberg A, Liñares J, et al. Pneumococcal capsular Nichol KA, et al. Assessment of multidrug resistance, clonality and switching: a historical perspective. J Infect Dis. 2013;207:439–49. virulence in non–PCV-13 Streptococcus pneumoniae serotypes in http://dx.doi.org/10.1093/infdis/jis703 Canada, 2011–13. J Antimicrob Chemother. 2015;70:1960–4. 26. Mavroidi A, Aanensen DM, Godoy D, Skovsted IC, Kaltoft MS, 12. Metcalf BJ, Chochua S, Gertz RE Jr, Li Z, Walker H, Tran T, et Reeves PR, et al. Genetic relatedness of the Streptococcus al.; Active Bacterial Core surveillance team. Using whole genome pneumoniae capsular biosynthetic loci. J Bacteriol. 2007; sequencing to identify resistance determinants and predict 189:7841–55. http://dx.doi.org/10.1128/JB.00836-07 antimicrobial resistance phenotypes for year 2015 invasive 27. Golubchik T, Brueggemann AB, Street T, Gertz RE Jr, Spencer CC, pneumococcal disease isolates recovered in the United States. Ho T, et al. Pneumococcal genome sequencing tracks a vaccine Clin Microbiol Infect. 2016;22:1002.e1–8. http://dx.doi.org/ escape variant formed through a multi-fragment recombination 10.1016/j.cmi.2016.08.001 event. Nat Genet. 2012;44:352–5. http://dx.doi.org/10.1038/ng.1072 13. Beall B, McEllistrem MC, Gertz RE Jr, Wedel S, Boxrud DJ, 28. Enright MC, Spratt BG. Extensive variation in the ddl gene of Gonzalez AL, et al.; Active Bacterial Core Surveillance Team. penicillin-resistant Streptococcus pneumoniae results from a Pre- and postvaccination clonal compositions of invasive hitchhiking effect driven by the penicillin-binding protein pneumococcal serotypes for isolates collected in the United States 2b gene. Mol Biol Evol. 1999;16:1687–95. http://dx.doi.org/ in 1999, 2001, and 2002. J Clin Microbiol. 2006;44:999–1017. 10.1093/oxfordjournals.molbev.a026082 http://dx.doi.org/10.1128/JCM.44.3.999-1017.2006 29. Kauffmann ,F Mørch E, Schmith K. On the serology of the 14. Beall BW, Gertz RE, Hulkower RL, Whitney CG, Moore MR, pneumococcus-group. J Immunol. 1940;39:397–426. Brueggemann AB. Shifting genetic structure of invasive serotype 30. Lund E. On the nomenclature of the pneumococcal types. 19A pneumococci in the United States. J Infect Dis. 2011;203: Int J Syst Bacteriol. 1970;20:321–3. http://dx.doi.org/10.1099/ 1360–8. http://dx.doi.org/10.1093/infdis/jir052 00207713-20-3-321 15. Li Y, Metcalf BJ, Chochua S, Li Z, Gertz RE Jr, Walker H, et al. 31. Henrichsen J. Six newly recognized types of Streptococcus Penicillin-binding protein transpeptidase signatures for tracking pneumoniae. J Clin Microbiol. 1995;33:2759–62. and predicting β-lactam resistance levels in Streptococcus 32. Kim L, McGee L, Tomczyk S, Beall B. Biological and pneumoniae. MBio. 2016;7:e00756–16. http://dx.doi.org/10.1128/ epidemiological features of antibiotic-resistant Streptococcus mBio.00756-16 pneumoniae in pre- and post-conjugate vaccine eras: a United 16. Clinical and Laboratory Standards Institute (CLSI). Performance States perspective. Clin Microbiol Rev. 2016;29:525–52. standards for antimicrobial susceptibility testing. Twenty-third http://dx.doi.org/10.1128/CMR.00058-15 informational supplement (M100–S22.CLSI). Wayne (PA): The 33. Olarte L, Kaplan SL, Barson WJ, Romero JR, Lin PL, Tan TQ, Institute; 2013. et al. Emergence of multidrug-resistant pneumococcal serotype 17. Martin M. Cutadapt removes adapter sequences from high- 35B among children in the United States. J Clin Microbiol. throughput sequencing reads. EMBnet. 2011;17:10–2. 2017;55:724–34. http://dx.doi.org/10.1128/JCM.01778-16 http://dx.doi.org/10.14806/ej.17.1.200 34. Moore MR, Gertz RE Jr, Woodbury RL, Barkocy-Gallagher GA, 18. Zerbino DR, Birney E. Velvet: algorithms for de novo short read Schaffner W, Lexau C, et al. Population snapshot of emergent assembly using de Bruijn graphs. Genome Res. 2008;18:821–9. Streptococcus pneumoniae serotype 19A in the United States, http://dx.doi.org/10.1101/gr.074492.107 2005. J Infect Dis. 2008;197:1016–27. http://dx.doi.org/10.1086/ 19. Gardner SN, Slezak T, Hall BG. kSNP3.0: SNP detection and 528996 phylogenetic analysis of genomes without genome alignment 35. Sobanjo-ter Meulen A, Vesikari T, Malacaman EA, Shapiro SA, or reference genome. Bioinformatics. 2015;31:2877–8. Dallas MJ, Hoover PA, et al. Safety, tolerability and http://dx.doi.org/10.1093/bioinformatics/btv271 immunogenicity of 15-valent pneumococcal conjugate vaccine in 20. Stamatakis A. RAxML version 8: a tool for phylogenetic toddlers previously vaccinated with 7-valent pneumococcal analysis and post-analysis of large phylogenies. Bioinformatics. conjugate vaccine. Pediatr Infect Dis J. 2015;34:186–94. 2014;30:1312–3. http://dx.doi.org/10.1093/bioinformatics/btu033 http://dx.doi.org/10.1097/INF.0000000000000516 21. Feil EJ, Li BC, Aanensen DM, Hanage WP, Spratt BG. eBURST: inferring patterns of evolutionary descent among clusters of related Address for correspondence: Bernard Beall, Centers for Disease Control bacterial genotypes from multilocus sequence typing data. and Prevention, 1600 Clifton Rd NE, Mailstop C02, Atlanta, GA J Bacteriol. 2004;186:1518–30. http://dx.doi.org/10.1128/ JB.186.5.1518-1530.2004 30329-4017, USA: email [email protected]

930 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Serologic and Molecular Evidence of Vaccinia Virus Circulation among Small Mammals from Different Biomes, Brazil Júlia B. Miranda, Iara A. Borges, Samantha P.S. Campos, Flávia N. Vieira, Tatiana M.F. de Ázara, Fernanda A. Marques, Galileu B. Costa, Ana Paula M.F. Luis, Jaqueline S. de Oliveira, Paulo César P. Ferreira, Cláudio Antônio Bonjardim, Silvio L.M. da Silva, Álvaro E. Eiras, Jônatas S. Abrahão, Erna G. Kroon, Betânia P. Drumond, Adriano P. Paglia, Giliane de S. Trindade

Vaccinia virus (VACV) is a zoonotic agent that causes a vaccination was discontinued (2), and only select institu- disease called bovine vaccinia, which is detected mainly in tions in the United States (e.g., the military and certain pub- milking cattle and humans in close contact with these ani- lic health facilities) receive the vaccine for their efforts to mals. Even though many aspects of VACV infection have prevent the use of VARV as a biologic weapon. Household been described, much is still unknown about its circulation transmission from vaccinees and eczema vaccinatum are in the environment and its natural hosts/reservoirs. To in- some of the negative aspects of vaccinating and have been vestigate the presence of Orthopoxvirus antibodies or VACV DNA, we captured small rodents and marsupials in 3 areas responsible for severe outcomes (3). Although VARV is of Minas Gerais state, Brazil, and tested their samples in now restricted to laboratory facilities, other OPVs have been a laboratory. A total of 336 animals were tested; positivity emerging as zoonotic pathogens in different geographic ar- ranged from 18.1% to 25.5% in the 3 studied regions lo- eas, namely CPXV in Europe, monkeypox virus (MPXV) in cated in different biomes, including the Atlantic Forest and Africa, and VACV in Asia and South America (4). the Cerrado. Analysis of nucleotide sequences indicated In Brazil, natural infections with VACV are called bo- co-circulation of VACV groups I and II. Our findings rein- vine vaccinia (BV) and are reported in rural areas, mainly force the possible role played by rodents and marsupials in in milking cattle and in men who are in close contact with VACV maintenance and its transmission chain. these animals. The first officially recorded reports of BV date from the early 2000s and occurred in the southeastern region irus species belonging to genus Orthopoxvirus (OPV) of Brazil (5,6). Currently, there is evidence of virus circula- Vreceive great attention because of Variola virus tion in all regions of Brazil (7,8); however, the southeast is (VARV), which is associated with smallpox (1). Smallpox still the epicenter of registered BV cases, with Minas Gerais caused many deaths worldwide and was eradicated after state being one of the most affected. Studies have shown that a massive vaccination campaign developed by the World mammal species in addition to bovids and humans could be Health Organization (2). Because OPVs have very similar naturally infected by (or at least exposed to) VACV (9–19). antigenic structure (1), cross-protection enabled the use of VACV was isolated from samples from a rodent from the cowpox virus (CPXV) and later vaccinia virus (VACV) as Amazon region in the 1960s (9), and now there are other anti-smallpox vaccine agents (2). documented incidents of virus circulation in these animals Given its widespread use, VACV has been studied (10–12,15,17). By taking into account virus detection in for many years, and these efforts shed light on various as- small rodents, the fact that CPXV (20) and probably MPXV pects regarding virus biology. After smallpox eradication, have rodents as reservoirs (21), and the frequent reports of these animals’ presence during BV outbreaks, an ecologic Author affiliations: Universidade Federal de Minas Gerais, model was created to propose the participation of rodents in Belo Horizonte, Brazil (J.B. Miranda, I.A. Borges, S.P.S. Campos, the VACV transmission chain (12). Because some species F.N. Vieira, T.M.F de Ázara, F.A. Marques, G.B. Costa, A.P.M.F. Luis, of native rodents could have ecologic advantages in areas J.S. de Oliveira, P.C.P. Ferreira, C.A. Bonjardim, A.E. Eiras, with anthropic disturbance, they could work as bridges be- J.S. Abrahão, E.G. Kroon, B.P. Drumond, A.P. Paglia, G.D.S. tween natural and human/domestic habitats, bringing viruses Trindade); Instituto Federal de Educação, Ciência e Tecnologia do from wild animals to domestic ones and vice versa (12). This Sudeste de Minas Gerais, Rio Pomba, Brazil (S.L.M. da Silva) model is reinforced by studies of virus transmission between mice and from experimental infection through contaminated DOI: https://dx.doi.org/10.3201/eid2306.161643

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 931 RESEARCH milk (22). To better evaluate the circulation of VACV in institution in a previous study (Figure 1, panel B). Three small rodents, we undertook comprehensive collection cam- sampling transects were demarcated: 1 in savannah veg- paigns in 3 areas of Minas Gerais with or without confirmed etation with intense anthropogenic disturbance and the BV outbreaks. Animals were evaluated for the presence of other 2 in forest vegetation (with 1 of the 2 having more VACV DNA and antibodies against OPV. Because marsupi- disturbance than the other) (Figure 1, panel B). In each als were often captured and previous studies have detected transect, 15 sampling points were established with 2 live OPV antibodies and VACV DNA in these animals (17,19), traps in each, 20 m apart, where captures of small mam- their samples were also tested. mal took place. In Rio Pomba, the field site (21°16′29′′S, 43°10′45′′W) Methods has characteristic Atlantic Forest vegetation. Animal trap- ping was performed in the area around the Instituto Feder- Collection Sites al de Educação, Ciência e Tecnologia (Figure 2, panels A, We collected small mammals in 3 areas of Minas Gerais. B). Animal collections were performed in forest, pasture, Brazil was chosen because of its history of BV outbreaks and peridomicile areas (Figure 2, panels A–G). In each and its different biomes and conservational status (Figure transect, 10 traps were placed at a distance of 10 m from 1, panel A). The 3 municipalities where collections were each other and in alternating positions (on forest floor or performed were Sabará, Serro, and Rio Pomba. on tree trunks). Sabará is a city located in an anthropic area situated Serro, a city whose economy is based on milk and in the transition from savannah (the Cerrado biome) to the cheese production, has seen many cases of BV since Atlantic Forest. The study site (19°53′9′′S, 43°48′45′′W) 2005 (23). The capture of small mammals was carried was delimited on the grounds of a former educational out in 2 farms (Figure 2, panels C, D). The study site (180°36′21.16′′S, 43°23′12.89′′W) is situated in the Cer- rado biome and has some intersections of Atlantic Forest. Animal collections were carried out in forest, pasture, and peridomicile areas (Figure 2, panels A–G), and traps were positioned in the same manner as in Rio Pomba.

Animal Trap and Sample Collection Captures lasted from April 2011 through May 2012 for Sabará (12 campaigns) and from September 2012 through September 2013 in Serro (5 campaigns) and Rio Pomba (6 campaigns). Small mammals were captured in size- selective live cages by using pineapple chunks and cot- ton balls soaked in cod liver oil as baits. Each sampling section lasted for 4 nights, and baits were replaced after 2 nights. After capture, animals were anesthetized with ketamine (70 mg/kg) and xylazine (12 mg/kg) for serum collection. Animals were weighed, measured for size, and visually evaluated for clinical signs of disease, such as skin lesions. For organ collection, animals were euthanized by intracardiac injection of 3 times the anesthetic dose according to guidelines of the American Society of Mammalogists (24). Collections were authorized by the Environment Ministry of Brazil through the SISBIO system (license no. 20807–2).

Figure 1. Locations of study areas, Minas Gerais state, Brazil. A) Biosafety Locations of the 3 municipalities where collections were performed: All collections were performed by trained professionals Sabará, Serro, and Rio Pomba. Inset shows location of Minas Gerais (either veterinarians or biologists) according to US Centers state in southeastern Brazil. B) Identification of 3 sample transects for Disease Control and Prevention recommendations (25). in Sabará. Trail 1 has savannah vegetation, and trails 2 and 3 have During animal manipulation, personal protective equip- Atlantic Forest vegetation. Sources: panel A, Scribble Maps; panel B, T.M.F. de Ázara. A color version of this figure is available online ment (disposable coveralls, surgical gloves, goggles, and (http://wwwnc.cdc.gov/EID/article/23/6/16-1643-F1.htm). N98 masks) was used.

932 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Vaccinia Virus among Small Mammals, Brazil

Figure 2. Location of collection sites and biomes represented in each, Minas Gerais state, Brazil. A) Collection site 1 in Serro. B) Collection site 2 in Serro. C) Peridomicile collection areas in Rio Pomba. D) Forest and pasture collection areas in Rio Pomba. E) Example of a forest area where animals were captured. F) Example of peridomicile area. G) Example of pasture area. In panels A–D, circles represent areas where transects for capture were demarcated. Sources: panels A,–D, Google Maps, modified by F.V. Nunes; panels E–G, F.V. Nunes. A color version of this figure is available online (http://wwwnc.cdc.gov/EID/article/23/6/16-1643-F2.htm).

DNA Extraction from Organs additionally tested for amplification of the A56R hemag- In addition to serum, which was tested by real-time PCR glutinin gene (primer sequences available upon request). (rPCR) targeting the C11R viral growth factor gene with- For C11R, an amplicon of 82 bp and a melting tempera- out DNA extraction, liver was the chosen organ for rPCR ture of 74.99°C were expected, and for A56R, a sequence trials. The organs were macerated with mortar and pestle of 160 bp and a melting temperature of 74.41°C were after liquid nitrogen was added, and DNA was extracted expected. All reactions had a final volume of 10 µL, and with PureLink Genomic DNA Mini Kit (Invitrogen, Carls- samples were tested in duplicates. Reaction steps com- bad, CA, USA) as recommended by the manufacturer. The prised initial DNA denaturation at 95°C for 10 min, 40 same protocol was applied for other organs tested, includ- cycles of denaturation (95°C for 15 s), annealing/exten- ing intestine, bladder, heart, gonads (ovary/testicles), bone sion (60°C for 60 s), and a melting curve (95°C for 15 s, marrow, spleen, lung, diaphragm, and kidney. 60°C for 60 s, and 95°C for 15 s). Samples were considered positive when melting temperature varied only up to Cells and Virus 1°C compared with a positive control (10 ng of DNA ex- A VACV Western Reserve strain was used as a positive tracted from purified VACV Western Reserve strain) and control. BSC-40 cells were grown in Eagle’s minimum had amplification in duplicate or for >1 target. Samples essential medium (Invitrogen) supplemented with 5% fe- with a single amplification were retested and considered tal bovine serum (Cultilab, São Paulo, Brazil); 25 mg/mL equivocal when no more amplification was observed. Fungizone (amphotericin B) (Cristália, São Paulo, Brazil); 500 U/mL penicillin; and 50 mg/mL gentamicin (Schering- Nucleotide Sequencing and Sequence Analyses Plough, São Paulo, Brazil). Positive samples that could be reamplified (A56R-positive) or amplified by a conventional PCR targeting C11R rPCR Assays (C11R-positive) (26) were chosen for sequencing. For All rPCR experiments were performed in 48-well plates A56R, product from the previous reaction was reamplified in Step One machines (Applied Biosystems, Foster City, in a conventional PCR reaction using 1 µL of the first reac- CA, USA) by using SYBR Green Master Mix (Applied tion as input and 0.2 nmol/L (A56R) rPCR primers. PCR Biosystems). DNA from liver samples was diluted in wa- cycling for A56R gene consisted of 10 min at 95°C for deter for a final use concentration of 10 ng/µL and 50 ng/ naturation, 30 cycles of denaturation (95°C for 10 min), µL. For serum samples, a 1:10 or 1:100 dilution was per- annealing (60°C for 60 s), extension (72°C for 60 s), and formed, and samples were tested without previous DNA a final extension of 10 min at 72°C. Products with single extraction. For both liver and serum samples, amplifica- bands were directly sequenced, and products with multiple tion of the C11R gene was tested, and liver samples were bands had the target gene extracted from acrylamide gels

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 933 RESEARCH stained with SYBR Gold Nucleic Acid Gel Stain (Invit- (C11R targeted in liver and serum and A56R in liver). One rogen) and had its DNA purified. Nucleotide sequencing animal was positive in the 3 tests performed, 11 in 2 tests, was performed by dideoxy method in an ABI3130 platform and 5 in only 1 test (data not shown). For the remaining (Applied Biosystems), and sequence quality was analyzed positive animals, >1 assays could not be performed, and 1 by using Sequence Scanner Software 1.0 (Applied Bio- sample type was positive in 1 test. systems). Sequences were aligned with other reference se- Of all the animals from the different collection sites in quences from the BLAST nucleotide database (http://blast. Sabará, 11/48 (22.9%) rodents and 3/76 (3.9%) marsupi- ncbi.nlm.nih.gov/Blast.cgi) by using MEGA 6.0; the same als were positive by rPCR. For Serro, no marsupials were program was used for identity matrix construction (27). positive but 4/25 (16.0%) rodents were positive. For Rio Pomba, 2/137 (1.4%) of rodents and 1/18 (5.5%) of mar- Plaque-Reduction Neutralization Test supials were positive (online Technical Appendix Table 1). The plaque-reduction neutralization test (PRNT) protocol Four rPCR-positive animals (2 rodents and 2 marsu- has been described previously by Geessien Kroon et al. pials) were selected for viral DNA detection in different (28). Samples were considered positive when a reduction organs by rPCR targeting the C11R and A56R genes. Posi- of >50% in virus plaque numbers was observed. tivity was found for heart, spleen, intestines, bladder, lungs, kidneys, and gonads (online Technical Appendix Table 2). ELISA No amplification was observed in any bone marrow or dia- ELISA was performed for rodent blood samples following phragm samples tested. a protocol also described previously (28). For each plate, 1 positive control (serum from Mus musculus experimen- OPV Antibodies in Serum from Free-living tally infected with VACV-Guarani P1) (29) and 3 negative Small Mammals Tested by PRNT controls (serum of noninfected M. musculus) were added. PRNT tests were performed in a total of 314 serum samples, Cutoff was established as the mean of negative controls op- and from these, 33 were considered positive, corresponding tical density units plus 3 times their SD. Samples with an to 10.5% of the animals. The reduction percentages var- optical density 10% above or below the cutoff were consid- ied from 50.5% to 95.6%. For the Sabará collection, posi- ered equivocal. tivity was 9.0% (10/111), 14.3% (6/42) for rodents and 5.8% (4/69) for marsupials. For Serro, positivity was 4.2% Interaction Networks (2/47), and only rodent samples were positive, correspond- Interaction networks are useful to help with understanding ing to 8.0% (2/25) of rodents tested. For Rio Pomba, posi- of virus-host dynamics. Each species is represented by a tivity was 13.4% (21/156), 14.3% (20/139) for rodents and vertex, and the link between 2 vertices represents the inter- 5.8% (1/17) for marsupials (online Technical Appendix action between 2 different species, making it possible to an- Table 1). alyze their interdependence (30). The networks also show which species have the higher number of positive samples OPV Antibodies in Serum from Free-living for virus detection and the area where they were collected. Small Mammals Tested by ELISA By using data from VACV-positive small mammal species, ELISA tests were performed on 189 rodent serum sam- we created weighted networks with the program Pajek 4.07 ples; a control serum for marsupials was not available. (31). Accordingly, adjacency matrices were generated for Of the animals tested, 19/189 (10.0%) were positive and each study area in which hosts were represented by lines 11/189 (5.8%) equivocal (online Technical Appendix and VACV by columns. Table 1). By location, 3/35 (8.5%) animals from Sabará, 9/25 (36.0%) from Serro, and 7/129 (5.4%) from Rio Results Pomba were positive. rPCR Amplification of VACV DNA from Free-living Sequencing Small Mammals Two PCR amplicons, amplified from Sabará animals, A total of 325 animals had their samples tested by rPCR were sequenced with C11R primers and resulted in se- targeting the C11R gene, the A56R gene, or both. Of these quences of 168 bp that were aligned with VACVs in Bra- animals, 21 (6.4%) tested positive (i.e., amplification in du- zil and other OPVs (online Technical Appendix Figure 1). plicates or in >1 sample/target) and 58 (17.8%) equivocal These 2 sequences had 100% similarity with each other; (online Technical Appendix Table 1, https://wwwnc.cdc. similarity with VACVs in Brazil ranged from 98.2% to gov/EID/article/23/6/16-1643-Techapp1.pdf). The cycle 100% and with CPXV from 87.3% to 89.1%, whereas thresholds varied from 28.42 to 39.33. From the total ani- similarity with VARV was 94.5% and with MPXV 90.9% mals tested by rPCR, 114 had samples available for all tests (data not shown). For A56R, sequencing was performed

934 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Vaccinia Virus among Small Mammals, Brazil in positive rPCR samples and resulted in 6 sequences of of positive samples in Serro was T. setosus (the hairy 102 bp. When compared with OPV sequences, 2 samples Atlantic spiny rat; Figure 3, panel B). For Rio Pomba, 5 from Sabará had an 18-nt deletion shared by Brazil VACV genera were positive, including Akodon sp. mice captured group I, whereas the other 4 samples from Sabará, Serro, in pasture and forest, Calomys tener and N. lasiurus mice and Rio Pomba did not have that deletion, being more captured in forest, and C. philander and Didelphis aurita similar to Brazil VACV group II and other OPVs (online opossums captured in pasture. The species with the high- Technical Appendix Figure 2). est number of positive animals in Rio Pomba was C. tener (the delicate vesper mouse), followed by N. lasiurus and Geographic and Species Distribution of Positivity Akodon sp. mice (Figure 3, panel C). For all areas studied, 336 animals belonging to 18 genera had their samples tested by rPCR, PRNT, and/or ELISA, Discussion and 65 (19.3%) were positive in >1 of these tests. Total In our study, we analyzed different biomes in an area positivity was 18.1% for Sabará, 25.5% for Serro, and where BV infections are common, and the positivity rates 18.5% for Rio Pomba (online Technical Appendix Table found for VACV were 25.7% for rodents and 7.6% for 1). A higher positivity was observed for rodents (25.7%) marsupials. Even though VACV is known to circulate in than for marsupials (7.6%) (data not shown). Brazil and cause a disease that leads to economic, social, Species identified among the test-positive rodents were and public health effects, few studies have been conducted Calomys sp., Akodon sp., Necromys lasiurus, Trinomys se- with the aim of clarifying the VACV transmission chain tosus, Cerradomys subflavus, Oligoryzomys sp., Nectomys and potential natural hosts (9–19,22,32). Previous stud- squamipes, Mus musculus, and Rattus rattus. For marsu- ies showed antibody positivity of 8.7%–17.9% for wild pials, the positive animals were characterized as species/ rodents captured in places with or without documented genera Didelphis sp., and Caluromys philander (online BV (15,17) and seropositivity of 8.2% for Didelphis spp. Technical Appendix Table 3). marsupials (17). Test-positive animals were captured in all sample ar- The higher positivity rate found for rodents in our eas in Sabará (savannah and forest) and Serro (pasture, study could be attributable to the use of 3 different tech- forest, and peridomicile areas), whereas test-positive ani- niques, including 2 techniques for detecting antibodies mals in Rio Pomba were captured in pasture and forest. (PRNT and ELISA) and 2 targets for DNA detection The interaction network for Sabará illustrates that 4 spe- (rPCR).The 3 techniques used to assess virus circulation cies had positive samples, with 3 of them found either in provide distinct responses about infection stages. Where- forest or savannah and the other in both areas. N. lasiurus as rPCR indicates the presence of viral DNA, the ELISA (the hairy-tailed bolo mouse) had the highest number of used in our study reveals the presence of IgG (indicative positive samples in Sabará (Figure 3, panel A). For Serro, of previous infection) (33), and PRNT detects neutral- evidence of VACV circulation was found in 6 species; izing antibodies that can be of different types, including of these, 2 were captured in forest, 2 in pasture, and 2 in IgG and IgM, which are produced early in the infection peridomicile areas. The species with the largest number process (34). Because we found positive animals for >1

Figure 3. Interaction networks for vaccina virus among small mammals in Sabará (A), Serro (B), and Rio Pomba (C) in Minas Gerais state, Brazil. The square represents vaccinia virus. Circles represent small mammal species (labeled). The color in the circles represents the area where mammals were collected. The thickness of lines increases with the number of positive samples from a species. Acokur, Akodon cursor mouse; Akomys, Akodon cf; mystax; Calexp, Calomys expulsus; Calphi, Caluromys philander; Calten, Calomys tener; Cersub, Cerradomys subflavus; Didalb, Didelphis albiventris; Didaur, Didelphis aurita; Mus, Mus musculus; Neclas, Necromys lasiurus; Necsqui, Nectomys squamipes; Olisp, Oligoryzomys sp.; Rattus, Rattus rattus; Triset, Trinomys setosus; VACV, vaccinia virus. A color version of this figure is available online (http://wwwnc.cdc.gov/EID/article/23/6/16-1643-F3.htm).

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 935 RESEARCH techniques, we can speculate that an active transmission opossums had already been found to be positive (17,19), cycle is happening in all 3 study areas. Additionally, and C. philander opossums also had positive samples. only Serro has recurrent reports of BV outbreaks (23,29), Although only 1 virus (VACV) was analyzed for in- which could be a result of the presence of positive animals teraction network construction and no interactions between in the peridomicile area, where they could infect other an- different species were observed, the networks created illus- imals, such as cows, and cause disease. Furthermore, be- trate the participation of the small mammals for each stud- cause large-scale milk production occurs in Serro, many ied area and the areas where these positive animals were livestock animals, including bovines, could work as infec- collected (Figure 3). The networks also suggest an impor- tion amplifiers. tant role of N. lasiurus mice for the VACV transmission The rPCR technique has been used for MPXV detec- chain in Sabará, T. setosus rats in Serro, and C. tener, N. tion in rodents in Africa, where samples were considered lasiurus, and Akodon sp. mice in Rio Pomba. Most of these equivocal when repeatability of results was not achieved species are generalist animals that can be adapted to a dis- (21). In our study, we made this same observation, which turbed environment. However, the C. philander opossum is might be attributable to a low virus load in the samples. In an arboreal species that lives in forests (40), which could turn, the low virus load could be related to the late cycle indicate that a wild cycle is being maintained and that other threshold in which amplification occurred and a lack of animals could be transporting the virus between forests and clinical signs in animals with a positive result. peridomicile areas. These findings corroborate the models Sequencing of A56R rPCR amplicons revealed the co- proposed by Abrahão et al. (12) in which rodents and other circulation of Brazil VACVs belonging to groups I and II, small mammals could work as links between natural and a fact that reinforces previous data on Brazil VACV virus anthropic environments. diversity (11,35–37). Again, even when infected with virus In conclusion, our findings reinforce evidence of par- belonging to group II, which were found to be virulent in ticipation of rodents and marsupials in the VACV trans- a mice model (38), wild rodents and marsupials tested in mission cycle and the possibility that these animals might our study did not have clinical signs detected. Also, these work as links between natural and anthropic environments. animals infected with VACV group I or II had viral DNA These findings also further illustrate the multi-host charac- in many organs, as indicated by rPCR. It is not possible to teristic of VACV infection in Brazil. assert that virus is replicating in these tissues, given that virus could be present in blood that circulates through these Acknowledgments organs; however, previous in vivo infection experiments We hank Mylleus and Plataforma de Sequenciamento have found virus in different mice organs (32), probably FIOCRUZ for sequencing. The VACV Western Reserve because of systemic infection. This observation also was strain was kindly provided by C. Jungwirth of Universitat made in mice infected with milk (a possible route of natu- Wurzburg, Germany. ral infection) contaminated with VACV-Guarani P2 virus. This virus was found to be nonvirulent in a mice model; Financial support was provided by Departamento de animals shed viral DNA and produced OPV antibodies Microbiologia, CNPq, CAPES, PRPq-UFMG, FAPEMIG, and but did not show clinical signs (22). The detection of vi- MAPA. G.S.T., E.G.K., A.E.E., J.S.A., and A.P. are researchers rus DNA in intestines, bladder, and gonads could reinforce from CNPq. previous data suggesting that virus transmission occurs Ms. Miranda is graduated in biological sciences from the through feces (39) and support the hypothesis of alterna- Universidade Federal de Minas Gerais and completed her tive transmission through urine and sexual contact. Mari- master’s degree at the same university working with ana virus has been isolated from the gonads of mice (14), emerging viral diseases. Her primary research interests are so it could be speculated that the sexual transmission route emerging diseases and viral ecology. is involved. Among the positive rodent species, Akodon sp., N. squamipes, Oligoryzomys sp. (15,17), and M. musculus References (12) have already been found to be positive in previous 1. Moss B. Poxviridae. In: Knipe DM, Howley PM, editors. Fields studies, reinforcing evidence of its participation in the virology, 6th ed. Philadelphia: Lippincott Williams & Wilkins; 2013. p. 2129–59. VACV transmission cycle; however, the exact role played 2. Fenner F, Wittek R, Dumbell KR. The global spread, control and by these animals is not yet known. In addition to M. mus- eradication of smallpox. In: The orthopoxviruses. San Diego (CA): culus, C. subflavus, N. lasiurus, T. setosus, C. tener, and Academic Press; 1989. p. 317–52. R. rattus rodents were also found to be positive in our 3. Vora S, Damon I, Fulginiti V, Weber SG, Kahana M, Stein SL, et al. Severe eczema vaccinatum in a household contact of study, indicating the role of multiple hosts in VACV trans- a smallpox vaccinee. Clin Infect Dis. 2008;46:1555–61. mission in Brazil. Among the marsupials, Didelphis spp. http://dx.doi.org/10.1086/587668

936 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Vaccinia Virus among Small Mammals, Brazil

4. Essbauer S, Pfeffer M, Meyer H. Zoonotic poxviruses. vaccinia virus during outbreak affecting cattle and humans, Vet Microbiol. 2010;140:229–36. http://dx.doi.org/10.1016/ São Paulo state, Brazil. Emerg Infect Dis. 2016;22:271–3. j.vetmic.2009.08.026 http://dx.doi.org/10.3201/eid2202.140747 5. Damaso CRA, Esposito JJ, Condit RC, Moussatché N. An emergent 20. McFadden G. Poxvirus tropism. Nat Rev Microbiol. 2005;3:201– poxvirus from humans and cattle in Rio de Janeiro state: Cantagalo 13. http://dx.doi.org/10.1038/nrmicro1099 virus may derive from Brazilian smallpox vaccine. Virology. 21. Reynolds MG, Carroll DS, Olson VA, Hughes C, Galley J, 2000;277:439–49. http://dx.doi.org/10.1006/viro.2000.0603 Likos A, et al. A silent enzootic of an orthopoxvirus in Ghana, West 6. de Souza Trindade G, da Fonseca FG, Marques JT, Nogueira ML, Africa: evidence for multi-species involvement in the absence of Mendes LCM, Borges AS, et al. Araçatuba virus: a vaccinialike widespread human disease. Am J Trop Med Hyg. 2010;82:746–54. virus associated with infection in humans and cattle. Emerg Infect http://dx.doi.org/10.4269/ajtmh.2010.09-0716 Dis. 2003;9:155–60. http://dx.doi.org/10.3201/eid0902.020244 22. Rehfeld IS, Guedes MIMC, Fraiha ALS, Costa AG, Matos ACD, 7. Brum MCS, Anjos BL, Nogueira CEW, Amaral LA, Weiblen R, Fiúza ATL, et al. Vaccinia virus transmission through Flores EF. An outbreak of orthopoxvirus-associated disease in experimentally contaminated milk using a murine model. horses in southern Brazil. J Vet Diagn Invest. 2010;22:143–7. PLoS One. 2015;10:e0127350. http://dx.doi.org/10.1371/ http://dx.doi.org/10.1177/104063871002200132 journal.pone.0127350 8. Oliveira DB, Assis FL, Ferreira PCP, Bonjardim CA, 23. Trindade GS, Guedes MIC, Drumond BP, Mota BEF, Abrahão JS, de Souza Trindade G, Kroon EG, et al. Group 1 Vaccinia virus Lobato ZIP, et al. Zoonotic vaccinia virus: clinical and zoonotic outbreak in Maranhao state, Brazil. Am J Trop Med Hyg. immunological characteristics in a naturally infected patient. Clin 2013;89:1142–5. http://dx.doi.org/10.4269/ajtmh.13-0369 Infect Dis. 2009;48:e37–40. http://dx.doi.org/10.1086/595856 9. Fonseca FG, Lanna MC, Campos MA, Kitajima EW, Peres JN, 24. Sikes RS, Gannon WH; Animal Care and Use Committee of the Golgher RR, et al. Morphological and molecular characterization American Society of Mammalogists. Guidelines of the American of the poxvirus BeAn 58058. Arch Virol. 1998;143:1171–86. Society of Mammalogists for the use of wild mammals in http://dx.doi.org/10.1007/s007050050365 research. J Mammal. 2011;92:235–53. http://dx.doi.org/10.1644/ 10. Diniz S, Trindade GS, Fonseca FG, Kroon EG. Surto de varíola 10-MAMM-F-355.1 murina em camundongos suíços em biotério—Relato de caso. 25. Mills JN, Yates TL, Childs JE, Parmenter RR, Ksiazek TG, Arquivo Brasileiro de Medicina Veterinária e Zootecnia. 2001; Rollin PE. Guidelines for working with rodents potentially infected 53:1–5. http://dx.doi.org/10.1590/S0102-09352001000200003 with hantavirus. J Mammal. 1995;76:716–22. http://dx.doi.org/ 11. Trindade GS, da Fonseca FG, Marques JT, Diniz S, Leite JA, 10.2307/1382742 De Bodt S, et al. Belo Horizonte virus: a vaccinia-like virus lacking 26. Abrahão JS, Drumond BP, Trindade GS, da Silva-Fernandes AT, the A-type inclusion body gene isolated from infected mice. J Gen Ferreira JMS, Alves PA, et al. Rapid detection of Orthopoxvirus by Virol. 2004;85:2015–21. http://dx.doi.org/10.1099/vir.0.79840-0 semi-nested PCR directly from clinical specimens: a useful 12. Abrahão JS, Guedes MICM, Trindade GS, Fonseca FG, Campos RK, alternative for routine laboratories. J Med Virol. 2010;82:692–9. Mota BF, et al. One more piece in the VACV ecological puzzle: http://dx.doi.org/10.1002/jmv.21617 could peridomestic rodents be the link between wildlife and 27. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: bovine vaccinia outbreaks in Brazil? PLoS One. 2009;4:e7428. Molecular Evolutionary Genetics Analysis version 6.0. Mol Biol http://dx.doi.org/10.1371/journal.pone.0007428 Evol. 2013;30:2725–9. http://dx.doi.org/10.1093/molbev/mst197 13. Abrahão JS, Silva-Fernandes AT, Lima LS, Campos RK, 28. Geessien Kroon E, Santos Abrahão J, de Souza Trindade G, Guedes MIMC, Cota MMG, et al. Vaccinia virus infection in Pereira Oliveira G, Moreira Franco Luiz AP, Barbosa Costa G, monkeys, Brazilian Amazon. Emerg Infect Dis. 2010;16:976–9. et al. Natural Vaccinia virus infection: diagnosis, isolation, http://dx.doi.org/10.3201/eid1606.091187 and characterization. Curr Protoc Microbiol. 2016;42:14A.5. 14. Campos RK, Brum MCS, Nogueira CEW, Drumond BP, Alves PA, 1–14A.5.43. PubMed https://doi.org/10.1002/cpmc.13 Siqueira-Lima L, et al. Assessing the variability of Brazilian 29. Assis FL, Borges IA, Ferreira PCP, Bonjardim CA, Trindade GS, Vaccinia virus isolates from a horse exanthematic lesion: Lobato ZIP, et al. Group 2 vaccinia virus, Brazil. Emerg Infect Dis. coinfection with distinct viruses. Arch Virol. 2011;156:275–83. 2012;18:2035–8. http://dx.doi.org/10.3201/eid1812.120145 http://dx.doi.org/10.1007/s00705-010-0857-z 30. Newman MEJ. The structure and function of complex networks. 15. Schatzmayr HG, Costa RVC, Gonçalves MCR, D’Andréa PS, Society for Industrial and Applied Mathematics Review. 2003; Barth OM. Human and animal infections by vaccinia-like viruses 45:167–256. in the state of Rio de Janeiro: a novel expanding zoonosis. 31. Batagelj V, Mrvar A. Pajek: program for large network analysis. Vaccine. 2011;29(Suppl 4):D65–9. http://dx.doi.org/10.1016/ Connections. 1998;21:47–57. j.vaccine.2011.09.105 32. Ferreira JMS, Abrahão JS, Drumond BP, Oliveira FM, Alves PA, 16. de Assis FL, Pereira G, Oliveira C, Rodrigues GOL, Cotta MMC, Pascoal-Xavier MA, et al. Vaccinia virus: shedding and horizontal Silva-Fernandes AT, et al. Serologic evidence of orthopoxvirus transmission in a murine model. J Gen Virol. 2008;89:2986–91. infection in buffaloes, Brazil. Emerg Infect Dis. 2012;18:698–700. http://dx.doi.org/10.1099/vir.0.2008/003947-0 https://doi.org/10.3201/eid1804.111800 33. Chaudhri G, Panchanathan V, Bluethmann H, Karupiah G. 17. Peres MG, Bacchiega TS, Appolinário CM, Vicente AF, Obligatory requirement for antibody in recovery from a primary Allendorf SD, Antunes JMAP, et al. Serological study of vaccinia poxvirus infection. J Virol. 2006;80:6339–44. http://dx.doi.org/ virus reservoirs in areas with and without official reports of 10.1128/JVI.00116-06 outbreaks in cattle and humans in São Paulo, Brazil. Arch Virol. 34. Buller RML, Palumbo GJ. Poxvirus pathogenesis. Microbiol Rev. 2013;158:2433–41. http://dx.doi.org/10.1007/s00705-013-1740-5 1991;55:80–122. 18. Barbosa AV, Medaglia MLG, Soares HS, Quixabeira-Santos JC, 35. Trindade GS, Lobato ZIP, Drumond BP, Leite JA, Trigueiro RC, Gennari SM, Damaso CR. Presence of neutralizing antibodies to Guedes MIMC, et al. Isolation of two vaccinia virus strains Orthopoxvirus in capybaras (Hydrochoerus hydrochaeris) in from a single bovine vaccinia outbreak in rural area from Brazil: Brazil. J Infect Dev Ctries. 2014;8:1646–9. http://dx.doi.org/ implications on the emergence of zoonotic orthopoxviruses. 10.3855/jidc.5216 Am J Trop Med Hyg. 2006;75:486–90. 19. Peres MG, Barros CB, Appolinário CM, Antunes JMAP, 36. Drumond BP, Leite JA, da Fonseca FG, Bonjardim CA, Mioni MSR, Bacchiega TS, et al. Dogs and opossums positive for Ferreira PCP, Kroon EG. Brazilian Vaccinia virus strains are

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 937 RESEARCH

genetically divergent and differ from theLister vaccine strain. of Vaccinia virus strains in murine feces: implications for Microbes Infect. 2008;10:185–97. http://dx.doi.org/10.1016/ virus circulation and environmental maintenance. Arch Virol. j.micinf.2007.11.005 2009;154:1551–3. http://dx.doi.org/10.1007/s00705-009-0470-1 37. Oliveira G, Assis F, Almeida G, Albarnaz J, Lima M, Andrade AC, 40. Paglia AP, Fonseca GAB, Rylands AB, Herrmann G, Aguiar LMS, et al. From lesions to viral clones: biological and molecular Chiarello AG, et al. Annotated checklist of Brazilian mammals. diversity amongst autochthonous Brazilian Vaccinia virus. Viruses. 2nd ed. Occasional papers in conservation biology series, no. 6. 2015;7:1218–37. http://dx.doi.org/10.3390/v7031218 Arlington (VA): Conservation International; 2012. p. 76. 38. Ferreira JMS, Drumond BP, Guedes MIMC, Pascoal-Xavier MA, Almeida-Leite CM, Arantes RME, et al. Virulence in murine Address for correspondence: Giliane de S. Trindade, Laboratório de model shows the existence of two distinct populations of Vírus, Instituto de Ciências Biológicas, sala 258, Universidade Brazilian Vaccinia virus strains. PLoS One. 2008;3:e3043. http://dx.doi.org/10.1371/journal.pone.0003043 Federal de Minas Gerais. Avenida Antônio Carlos, 6627 - Pampulha 39. Abrahão JS, Trindade GS, Ferreira JMS, Campos RK, - Belo Horizonte, Minas Gerais, Brazil, CEP 31270-901; email: Bonjardim CA, Ferreira PCP, et al. Long-lasting stability [email protected]

December 2015: Zoonotic Infections

• Biological Warfare Plan in the 17th Century—the Siege of Candia, 1648–1669 • Influenza A(H6N1) Virus in Dogs, Taiwan • Methicillin-Resistant Staphylococcus aureus Prevalence among Captive Chimpanzees, Texas, • Increased Number of Human USA, 2012 Cases of Influenza Virus • Novel Waddlia Intracellular A(H5N1) Infection, • Identifying and Reducing Bacterium in Artibeus Egypt, 2014–15 Remaining Stocks of intermedius Fruit Bats, Mexico • Replication Capacity of Avian Rinderpest Virus • Tembusu-Related Flavivirus in Influenza A(H9N2) Virus in Pet • Opportunistic Pulmonary Ducks, Thailand Birds, Chickens, and Mammals, Bordetella hinzii Infection after Bangladesh Avian Exposure • Hendra Virus Infection in Dog, • Zoonotic Leprosy in the Australia, 2013 Southeastern United States • No Evidence of Gouléako and • Infection Risk for Persons Herbert Virus Infections in Pigs, Exposed to Highly Pathogenic Côte d’Ivoire and Ghana Avian Influenza A H5 • Aquatic Bird Bornavirus 1 in Virus–Infected Birds, United Wild Geese, Denmark States, December 2014– March 2015 • Japanese Macaques (Macaca • Vectorborne Transmission of fuscata) as Natural Reservoir of Leishmania infantum from • High Prevalence of Intermediate Bartonella quintana Hounds, United States Leptospira spp. DNA in Febrile Humans From Urban and • Onchocerca lupi Nematode in • Porcine Deltacoronavirus in Rural Ecuador a Cat, Europe Mainland China

https://wwwnc.cdc.gov/eid/articles/ ® issue/21/12/table-of-contents

938 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Relative Risk for Ehrlichiosis and Lyme Disease in an Area Where Vectors for Both Are Sympatric, New Jersey, USA Andrea Egizi, Nina H. Fefferman, Robert A. Jordan

In support of improving patient care, this activity has been planned and implemented by Medscape, LLC and Emerging Infectious Diseases. Medscape, LLC is jointly accredited by the Accreditation Council for Continuing Medical Education (ACCME), the Accreditation Council for Pharmacy Education (ACPE), and the American Nurses Credentialing Center (ANCC), to provide continuing education for the healthcare team. Medscape, LLC designates this Journal-based CME activity for a maximum of 1.00 AMA PRA Category 1 Credit(s)™. Physicians should claim only the credit commensurate with the extent of their participation in the activity. All other clinicians completing this activity will be issued a certificate of participation. To participate in this journal CME activity: (1) review the learning objectives and author disclosures; (2) study the education content; (3) take the post- test with a 75% minimum passing score and complete the evaluation at http://www.medscape.org/journal/eid; and (4) view/print certificate. For CME questions, see page 1,056. Release date: May 11, 2017; Expiration date: May 11, 2018

Learning Objectives

Upon completion of this activity, participants will be able to: • Analyze the clinical presentation of ehrlichiosis. • Compare the vectors of ehrlichiosis vs Lyme disease. • Distinguish the ratio of ehrlichiosis to Lyme disease using a mathematical model. • Compare predicted rates of ehrlichiosis with actual reported rates of illness. CME Editor Jude Rutledge, BA, Technical Writer/Editor, Emerging Infectious Diseases. Disclosure: Jude Rutledge has disclosed no relevant financial relationships. CME Author Charles P. Vega, MD, Health Sciences Clinical Professor, UC Irvine Department of Family Medicine; Associate Dean for Diversity and Inclusion, UC Irvine School of Medicine, Irvine, California, USA. Disclosure: Charles P. Vega, MD, has disclosed the following financial relationships: served as an advisor or consultant for McNeil Consumer Healthcare; served as a speaker or a member of a speakers bureau for Shire Pharmaceuticals. Authors Disclosures: Andrea Egizi, PhD; and Robert A. Jordan, PhD, have disclosed no relevant financial relationships. Nina H. Fefferman, PhD, has disclosed the following relevant financial relationships: owns stock, stock options, or bonds from VIVUS, Inc.

Author affiliations: Rutgers University, New Brunswick, The lone star tick, Amblyomma americanum, is a vector of New Jersey, USA (A. Egizi, N.H. Fefferman, R.A. Jordan); Ehrlichia chaffeensis and E. ewingii, causal agents of hu- Monmouth County Mosquito Control Division, Tinton Falls, man ehrlichiosis, and has demonstrated marked geograph- New Jersey, USA (A. Egizi, R.A. Jordan); University of Tennessee, ic expansion in recent years. A. americanum ticks often Knoxville, Tennessee, USA (N.H. Fefferman) outnumber the vector of Lyme disease, Ixodes scapularis, where both ticks are sympatric, yet cases of Lyme disease DOI: http://dx.doi.org/10.3201/eid2306.160528 far exceed ehrlichiosis cases. We quantified the risk for

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 939

Page 1 of 1 RESEARCH ehrlichiosis relative to Lyme disease by using relative tick years, submissions of A. americanum ticks to the Monmouth encounter frequencies and infection rates for these 2 spe- County Mosquito Control Division’s passive tick surveil- cies in Monmouth County, New Jersey, USA. Our calcula- lance program (Tinton Falls, New Jersey, USA; offering free tions predict >1 ehrlichiosis case for every 2 Lyme disease tick identification to residents since 2006) have been increas- cases, >2 orders of magnitude higher than current case ing steadily while submissions of I. scapularis ticks remain rates (e.g., 2 ehrlichiosis versus 439 Lyme disease cases nearly level (Figure 1), suggesting increased human expo- in 2014). This result implies ehrlichiosis is grossly underreported (or misreported) or that many infections are asymp- sure to A. americanum ticks relative to I. scapularis ticks. tomatic. We recommend expansion of tickborne disease A. americanum ticks are known vectors of >2 human education in the Northeast United States to include human pathogens, Ehrlichia chaffeensis and E. ewingii (10,11), health risks posed by A. americanum ticks. which cause human ehrlichiosis, an illness often marked by an initial prodrome of undifferentiated fever, headache, myalgia, nausea, malaise, thrombocytopenia, leukopenia, ickborne diseases are a growing public health concern and hepatic injury (elevated serum transaminase levels) Tin the United States (1). Lyme disease, caused by the (12). More severe complications (including toxic shock– bacterium Borrelia burgdorferi, is the most frequently re- like symptoms and meningitis or meningoencephalitis) ported vectorborne illness in the Northeast (2) and has been may occur in some untreated persons (including immu- the subject of widespread education campaigns aimed at nocompromised patients), and death rates as high as ≈3% preventing human encounters with its vector, Ixodes scapu- have been reported (12,13). laris (the black-legged tick). Unfortunately, such campaigns The number of ehrlichiosis cases attributable to E. focus comparatively less attention on other medically im- chaffeensis that have been reported to the Centers for Dis- portant ticks in Lyme disease–endemic areas (3); therefore, ease Control and Prevention (Atlanta, GA, USA) has in- persons living in these areas may not fully recognize the creased steadily since the disease became reportable, from threat posed by these species. Specifically, the much more 0.8 cases/million persons/year in 2000 to 3.0 cases/million aggressive lone star tick, Amblyomma americanum, trans- persons/year in 2007 (13,14). However, some evidence mits the agent of human monocytic ehrlichiosis and may suggests that ehrlichiosis may be substantially underre- serve as the vector for several other emerging tickborne ported (12) and even reported cases may be misclassified pathogens (3–5). Historically found primarily in the South- because of cross-reactivity between E. chaffeensis, E. ew- east United States but with established distributions along ingii, and other Ehrlichia and Anaplasma agents in widely the Atlantic Coast and into the Midwest (6), A. america- used serologic tests (15–17). Health risks may be com- num ticks are active concurrently with I. scapularis ticks; pounded if physicians are less familiar with ehrlichiosis aggressively attack humans in all tick life stages (adult, than Lyme disease, particularly because initial symptoms nymph, and larvae); and are typically by far the more nu- may be relatively vague, resembling a viral syndrome typi- merous of the 2 ticks where they are sympatric (7,8). cal of an array of tickborne diseases (14), which creates the Given the abundance and aggressive host-seeking be- potential for diagnostic confusion where I. scapularis and havior of A. americanum ticks, it is reasonable to expect A. americanum ticks are sympatric (10). Ehrlichiosis could high rates of human encounters with them (8–10). In recent also be misdiagnosed as Rocky Mountain spotted fever if a patient is co-infected with Ehrlichia sp. and Rickettsia amblyommatis, (formerly Candidatus Rickettsia amblyommii [18]), a species cross-reactive in tests for rickettsial pathogens (19) and commonly present in >40% of field-collected A. americanum ticks (4). This study quantified the risk to humans of ehrlichial infections, relative to Lyme disease risk. We used data obtained from tick surveillance programs in Monmouth County, New Jersey, an area with a high reported incidence of Lyme disease and increasing A. americanum tick encounter rates.

Methods

Number of ticks submitted each year to Monmouth Figure 1. Site Description County Mosquito Control Division’s passive surveillance program ° ° during May–August, by year, Monmouth County, New Jersey, Monmouth County (40 44′N, 74 17′W) is located in USA, 2006–2015. eastern-central New Jersey, a US state on the mid-Atlantic

940 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Ehrlichiosis and Lyme Disease, New Jersey, USA coast. The county is 468.8 mi2 in size and had a population of surveillance data in our model (where A. americanum 630,380 (1,344.7 persons/mi2) as of the 2010 census (http:// ticks are only encountered 1.5 times as often; Table) pro- www.census.gov/quickfacts/table/PST045215/34025). vides a more direct measure of human exposure to ticks as The geomorphologic break separating the Inner and Outer well as a more conservative risk estimate. Coastal Plain physiographic provinces in New Jersey runs To characterize risk for exposure to the disease from horizontally across the center of the county, and the result- the vector, we define Ix as the prevalence of the pathogen ing soil differences are reflected in vegetative differences in ticks (i.e., percentage infected), weighted by life stage. between these 2 regions (20). The Outer Coastal Plain re- Both I. scapularis and A. americanum ticks have a 3-host gion is characterized by sandier soils that are often dry and life cycle, meaning that adults have had more opportuni- acidic, with pine forests and cedar swamps. Previously, the ties to feed on an infected host than nymphs and conse- distribution of A. americanum ticks in Monmouth Coun- quently infection rates differ between life stages. Because ty was restricted to this southern part of the county (21), transovarial transmission of either B. burgdorferi or E. whereas I. scapularis ticks were found throughout. How- chaffeensis does not occur (23,24), host-seeking larvae ever, recently specimens of A. americanum ticks have been are not infected and therefore were not included in the captured in the far north and west of the county. calculations. Relative abundance of nymphs and adults The county reports several hundred cases of Lyme of each species submitted to our passive tick surveil- disease annually, with a 10-year average (during 2005– lance program during May–August were reported (CX,N

2014) of 361 cases/year. By contrast, during that same and CX,D) and used to weight the infection prevalence of period, there were on average 5.5 cases/year of E. each (Table). Infection rates of I. scapularis ticks with B. chaffeensis infection and no cases of infection attributed burgdorferi for both life stages (IIS,D and IIS,N) also were to E.ewingii (22). obtained from our passive surveillance program, whereby residents submitting an I. scapularis tick can elect to have Risk Model it tested for B. burgdorferi through nested PCR assay (fol- The relative risk for infection with a tickborne pathogen lowing established protocols [8]). Our records show that depends on multiple factors, including risk for exposure during a 10-year period of our program, 90.5% of resi- to a competent vector, risk for exposure to the pathogen dents submitting I. scapularis ticks during May–August from exposure to the vector, and risk for transmission have chosen to have them tested (R.A. Jordan, unpub. from exposure to a pathogen-infected vector. To charac- data), including 153 adults and 1,146 nymphs. However, terize the risk for human exposure to ticks, we define the the program does not test A. americanum ticks, so infec- parameter Cx as the relative proportion of each species tion rates for this species were obtained from other sourc- (x) of tick submitted to the Monmouth County Mosquito es. Rates of adult tick infection with E. chaffeensis and E.

Control Division’s tick identification and testing service. ewingii (IAA,D) in Monmouth County were derived from This passive surveillance program, initiated in 2006, al- Schulze et al. (21) and summed, yielding a total value (ac- lows county residents to submit ticks they have encoun- counting for co-infected ticks) of 11.7% infection with tered (e.g., found on their skin or clothing) for species human Ehrlichia pathogens (N = 291). Nymphal infection identification. This program averages 658.9 submissions/ rates with both ehrlichia species were obtained from an year, although this number has increased markedly in re- unpublished dataset consisting of field-collected nymphs cent years (R.A. Jordan, unpub. data). For Cx, we used the from 4 sites in eastern and western Monmouth County 10-year average of relative submissions data (2006–2015) in 2014 (R.A. Jordan unpub. data). Nymphal specimens during peak Lyme disease transmission season in New were disrupted by using a TissueLyser and DNA isolat- Jersey (May–August) (Table). Although in Monmouth ed with QIAgen DNeasy 96 well-plate blood and tissue A. americanum ticks are often 3 times as abundant as I. kits (QIAGEN, Valencia, CA, USA). Specimens were scapularis ticks in field collections 9( ), using the passive tested for pathogens by using real-time PCR protocols for

Table. Parameters used in relative risk calculations for ehrlichiosis and Lyme disease, by vector, Monmouth County, New Jersey, USA* Parameter Ixodes scapularis Amblyomma americanum Relative abundance of each species CIS = 38.32 CAA = 61.68 Adults Nymphs Adults Nymphs Relative abundance of each life stage CIS. D = 19.96 CIS. N = 80.04 CAA. D = 35.34 CAA. N = 64.66 Infection rates per life stage IIS. D = 39.87 IIS. N = 23.3 IAA. D = 11.7 IAA. N = 9.04 Infection rates, weighted IIS = 26.60 IAA = 9.98 *Values are in percentages. Relative abundances (denoted by CX) are derived from specimens submitted to the Monmouth County Mosquito Control Division’s tick identification and testing service during peak Lyme disease transmission season (May–August) and during a 10-year period (2006–2015). Infection rates of I. scapularis ticks with Borrelia burgdorferi (IIS) also from passive surveillance program. Infection rates of A. americanum ticks (IAA) encompass both Ehrlichia chaffeensis and E. ewingii (accounting for co-infection).

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 941

RESEARCH

E. chaffeensis (25) and E. ewingii (26). Both probes were unknown. Therefore, although we are including the pa- modified slightly (shortened) to allow the use of an MGB rameter TX in the equation so that it can be applied when quencher as follows (5′–3′): VIC-CGGACAATTGCT- such data become better known, for purposes of these TATAACC-MGBNFQ for E. chaffeensis and 6FAM- analyses we are setting it equal to 1 for both diseases, AACAATTCCTAAATAGTCTCTGAC-MGBNFQ for yielding no functional impact on the model. E. ewingii. Sequence detection primers and reaction conditions were as described previously (26). A subset of Calculations samples was compared with conventional PCR methods Based on our definitions, we calculate the relative risk for × × established by this laboratory for detection of these patho- ehrlichiosis compared to Lyme disease as risk = (CAA IAA × × gens (21), and the 2 methods were found to be in agree- TAA)/(CIS IIS TIS). To then translate this value into expect- ment. The resulting infection prevalence of Ehrlichia ed cases of observed human infection, we move from risk to pathogens in A. americanum nymphs (IAA,N) (encompass- reported case numbers by using the reported number of cas- ing E. chaffeensis and E. ewingii, accounting for coinfec- es of Lyme disease in Monmouth County as a benchmark. tion) was 9.04% (N = 752). Overall infection prevalence In other words, if there were X Lyme disease cases, then

(Ix) in each species of tick, weighted by life stage, was the expected number of ehrlichiosis cases would be X mul- calculated by multiplying the relative abundance of each tiplied by the relative risk estimate calculated as described. life stage times its infection rate and summing across life stages (Table). Results

To characterize transmission risk, we defined TX as The relative risk for ehrlichiosis cases compared to Lyme the likelihood of successful pathogen transmission from disease was calculated as risk = (61.68 × 9.98 × 1)/(38.32 an infected tick to a human. The general concept of vec- × 26.61 × 1) = 0.604. These numbers mean that we should tor competence includes both this direction of transmis- expect to see ehrlichiosis cases occur 0.604 times as often sion from vector to host as well as the probability of as Lyme disease cases. the vector becoming infected from feeding on infected In 2014, a total of 439 cases of Lyme disease were hosts, which in our calculations is already reflected in reported in Monmouth County (22). By using the risk esti- tick infection rates (Ix). Data on transmission efficiency mates described, we would expect there to be >265 cases of to animals for both species is scarce because most stud- ehrlichiosis, >2 orders of magnitude higher than the num- ies feed multiple infected ticks on 1 host (so the risk ber of cases actually observed (Figure 2). imposed by a single feeding tick is unknown) and, because of the difficulty in working with large vertebrates Discussion in a laboratory setting, have very small sample sizes We demonstrate that in Monmouth County, New Jersey, (27–30). Further, whether probabilities of transmission ehrlichiosis infections from A. americanum ticks should obtained from animal studies are applicable to humans is be occurring, at a minimum, one half as often as Lyme disease (e.g., 1 ehrlichiosis case for every 2 Lyme disease cases). This rate of occurance is clearly not the case (Fig- ure 2), and these numbers suggest that ≈99% of potential ehrlichiosis infections are not recognized. It is possible that not all persons who become infected in Monmouth are county residents (and so their case would be recorded elsewhere) making the total number of Monmouth Coun- ty–derived infections likely to be somewhat higher than the observed 2 cases. However, even if one assumes all ehrlichiosis reported for the entire state of New Jersey to originate in Monmouth County (64 cases in 2014) (22) these values still indicate a substantial discrepancy between numbers of observed and expected cases. When selecting values for the parameters used in our calculations, every opportunity to be conservative was taken to avoid biasing estimates of relative risk. For example,

infection (Ix) probabilities were higher for Lyme disease Figure 2. Number of observed versus expected ehrlichiosis than ehrlichiosis (39.87% vs. 11.7% for adult ticks and cases, Monmouth County, New Jersey, USA, 2014. Expected values calculated by using number of observed Lyme disease 23.3% vs. 9.04% for nymphs). Although reported B. burg- cases as benchmark. dorferi infection rates in I. scapularis adults frequently

942 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Ehrlichiosis and Lyme Disease, New Jersey, USA range from 40% to 50% in hyperendemic areas, several samples from children from the Southeast United States for studies have reported lower rates (9,31). Our weighted in- E. chaffeensis revealed that many more children had been fection prevalence estimate of 9.98% (encompassing both exposed than showed clinical signs (41). Most documented E. chaffeensis and E. ewingii across adults and nymphs) cases of E. chaffeensis and E. ewingii infection come from is probably lower than the actual value, given that many older adults (13,15), so when younger adults and children studies have reported infection prevalence in the range of become infected, they may be less likely to have symptoms 5%–15% for E. chaffeensis alone and in some locations or seek treatment (41), accounting for the large number of twice that rate (3,32). Last, use of passive surveillance unreported infections. Other vectorborne diseases, such numbers (Cx) probably underestimates the actual risk for as West Nile virus, transmitted by mosquitoes, also dem- exposure to A. americanum ticks; residents who recognize onstrate large numbers of asymptomatic infections and the tick species may be less likely to bring in A. america- increased severity among the older population (42). How- num ticks to the passive surveillance program, because we ever, because of cross-reactivity in the serologic test for E. only test I. scapularis (as stated on our website), and taking chaffeensis and the lack of specific testing for E. ewingii, the 10-year average for relative abundances (2006–2015) one could argue that many presumed asymptomatic cases does not account for the recent surge in A. americanum tick of E. chaffeensis were actually incidences of infection with submissions from 2012 onward (Figure 1). In light of these E. ewingii, which tends to be more mild (16). Although our considerations, the actual risk for A. americanum tick–as- study considered E. chaffeensis and E. ewingii interchange- sociated ehrlichioses in Monmouth County may be much ably, if we repeat our calculations by using E. chaffeensis higher than we have estimated. infection rates alone, we would expect Monmouth County One caveat should be noted. Differences in PCR sensi- to have seen 84 cases of E. chaffeensis in 2014, when only 2 tivity between B. burgdorferi and Ehrlichia sp. assays could were reported. Therefore, 97.6% of E. chaffeensis infections affect our ability to compare infection rates, although to mit- are potentially going unnoticed (versus 99% of ehrlichiosis igate this problem as much as possible, we relied on estab- infections overall), which is still a troubling discrepancy. lished primers and checked that infection rates were within Across the entire United States, the number of Lyme ranges reported by other studies as described previously. disease cases dwarfs ehrlichiosis cases. During 2004–2013, Thus, any difference in our ability to detect pathogens be- the Centers for Disease Control and Prevention annually tween the 2 tick species is unlikely to alter our conclusions. reported 279–528 Lyme disease cases/1 million persons One possible explanation for the lower-than-expected (2). In contrast, during that same period, annual ehrlichiosis number of reported ehrlichiosis cases is a lack of aware- cases ranged from 3.3 to 26 cases/1 million persons (43). If ness about ehrlichial disease on the part of the public and our numbers can be extrapolated to other areas in the coun- physicians, leading to misdiagnosis and underreporting. try, this implies that, on a national scale, potentially thou- The infection tends to manifest as a general influenza-like sands of ehrlichiosis cases are going undiagnosed. The A. illness, and onset of a rash is rare (12), so persons may be americanum tick appears to be expanding into geographic less likely to visit a doctor unless more severe symptoms areas where it did not occur previously (6,32) and is be- emerge or they are aware of a recent tick bite and the pres- coming more abundant within its existing range (9,21,44). ence of tickborne diseases in the area. Awareness of non– Because ehrlichiosis cases have steadily increased since Lyme disease tickborne illnesses is startlingly low, even becoming reportable (13,15), the spread of A. americanum in parts of the country where ehrlichiosis cases outnumber ticks and the emergence of ehrlichiosis as a human patho- Lyme disease cases (33–35). One study found that >50% gen in the United States may parallel increases in I. scapu- of respondents in the United States had heard of Lyme dis- laris tick populations and the emergence of Lyme disease ease, whereas only 1.4% had heard of ehrlichiosis (35). As that occurred 30 years prior (14). Even if most unrecog- a consequence of these factors, when active screening for nized infections are mild or asymptomatic, these could still ehrlichial infections is performed, the resulting case rates have consequences for public health; for example, blood are often much higher than those reported to governmental donors who are unknowingly infected could pass the infec- agencies (3,36,37). For example, Olano et al. (37) found tion to immunocompromised patients (45), or prescription that the incidence of ehrlichiosis observed when actively of sulfa drugs for unrelated ailments could result in wors- screening patients was as much as 2 orders of magnitude ened disease presentation (46). higher than the passively reported incidence. Our findings indicate a need to increase public- edu Another explanation could be the existence of asymp- cation efforts about the risks for acquiring tickborne- dis tomatic infections. Several studies of E. chaffeensis anti- eases other than Lyme disease in the United States, and in body seroprevalence in adults found that most of those particular, to expand prevention awareness of medically carrying the antibodies had no recollection of a symptom- important tick species other than I. scapularis. Fortunate- atic infection (38–40). Further, a study screening blood ly, many of these diseases, including ehrlichiosis, can be

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 943 RESEARCH prevented in the same manner as Lyme disease (e.g., avoid- 4. Mixson TR, Campbell SR, Gill JS, Ginsberg HS, Reichard MV, ance of tick bites) and are treated similarly to Lyme disease Schulze TL, et al. Prevalence of Ehrlichia, Borrelia, and Rickettsial agents in Amblyomma americanum (Acari: Ixodidae) collected (12,47); also, the number of ehrlichiosis cases peaks during from nine states. J Med Entomol. 2006;43:1261–8. spring and summer months, corresponding with peak Lyme 5. Apperson CS, Engber B, Nicholson WL, Mead DG, Engel J, disease transmission months (13,15). Consequently, refocus- Yabsley MJ, et al. Tick-borne diseases in North Carolina: is ing existing public health education efforts to encompass the “Rickettsia amblyommii” a possible cause of rickettsiosis reported as Rocky Mountain spotted fever? Vector Borne Zoonotic Dis. full spectrum of tickborne diseases could be accomplished 2008;8:597–606. http://dx.doi.org/10.1089/vbz.2007.0271 without changing much of its content, although some attention 6. Springer YP, Eisen L, Beati L, James AM, Eisen RJ. Spatial should be given to the risks imposed by viruses and protozoa distribution of counties in the continental United States with records as well as unique characteristics of the questing behavior of of occurrence of Amblyomma americanum (Ixodida: Ixodidae). J Med Entomol. 2014;51:342–51. http://dx.doi.org/10.1603/ME13115 A. americanum ticks. To better inform these educational ef- 7. Stromdahl EY, Randolph MP, O’Brien JJ, Gutierrez AG. Ehrlichia forts and more accurately assess tickborne disease risk, more chaffeensis (Rickettsiales: Ehrlichieae) infection in Amblyomma research into diseases other than Lyme disease is required. americanum (Acari: Ixodidae) at Aberdeen Proving Ground, Humans can alter their environment in many ways that Maryland. J Med Entomol. 2000;37:349–56. http://dx.doi.org/ 10.1093/jmedent/37.3.349 affect disease transmission, from localized changes affecting 8. Schulze TL, Jordan RA, Schulze CJ, Mixson T, Papero M. tick habitat and host abundance, to larger changes affecting Relative encounter frequencies and prevalence of selected Borrelia, the planet’s climate. As these changes continue to occur, the Ehrlichia, and Anaplasma infections in Amblyomma americanum study of vector and pathogen distributions and abundance and Ixodes scapularis (Acari: Ixodidae) ticks from central New Jersey. J Med Entomol. 2005;42:450–6. will be critical to understanding the potential risk to humans 9. Schulze TL, Jordan RA, Healy SP, Roegner VE, Meddis M, posed by emerging pathogens. Our calculations imply that Jahn MB, et al. Relative abundance and prevalence of selected infections with E. chaffeensis and E. ewingii are underrec- Borrelia infections in Ixodes scapularis and Amblyomma ognized, at least in Monmouth County, New Jersey, if not americanum (Acari: Ixodidae) from publicly owned lands in Monmouth County, New Jersey. J Med Entomol. 2006;43:1269–75. throughout a larger portion of the United States where A. http://dx.doi.org/10.1093/jmedent/43.6.1269 americanum ticks are abundant or becoming abundant. Ad- 10. Stromdahl EY, Hickling GJ. Beyond Lyme: aetiology of tick-borne ditional effort is needed to determine the causes for this ap- human diseases with emphasis on the south-eastern United States. parent discrepancy and to characterize the actual prevalence Zoonoses Public Health. 2012;59(Suppl 2):48–64. http://dx.doi.org/10.1111/j.1863-2378.2012.01475.x of ehrlichiosis in the human population, as well as to raise 11. Goddard J, Varela-Stokes AS. Role of the lone star tick, Amblyomma awareness about the risk for exposure to these pathogens in americanum (L.), in human and animal diseases. Vet Parasitol. areas where A. americanum ticks are common. 2009;160:1–12. http://dx.doi.org/10.1016/j.vetpar.2008.10.089 12. Dumler JS, Madigan JE, Pusterla N, Bakken JS. Ehrlichioses in humans: epidemiology, clinical presentation, diagnosis, Acknowledgments and treatment. Clin Infect Dis. 2007;45(Suppl 1):S45–51. The authors would like to acknowledge Terry L. Schulze and http://dx.doi.org/10.1086/518146 Holly Gaff for insightful discussions, Vivien E. Roegner for her 13. Dahlgren FS, Mandel EJ, Krebs JW, Massung RF, McQuiston JH. Increasing incidence of Ehrlichia chaffeensis and Anaplasma role in laboratory analyses, and the comments of 2 anonymous phagocytophilum in the United States, 2000–2007. reviewers that greatly improved this manuscript. Am J Trop Med Hyg. 2011;85:124–31. http://dx.doi.org/10.4269/ ajtmh.2011.10-0613 Dr. Egizi oversees a research laboratory for the Monmouth 14. Paddock CD, Childs JE. Ehrlichia chaffeensis: a prototypical County Mosquito Control Division located in the Center for emerging pathogen. Clin Microbiol Rev. 2003;16:37–64. Vector Biology at Rutgers University in New Brunswick, http://dx.doi.org/10.1128/CMR.16.1.37-64.2003 15. Nichols Heitman K, Dahlgren FS, Drexler NA, Massung RF, New Jersey, USA. She is interested in using molecular tools to Behravesh CB. Increasing Incidence of ehrlichiosis in the United gain insight into the ecology and evolution of vectors and States: a summary of national surveillance of Ehrlichia chaffeensis vectorborne disease in the northeastern United States. and Ehrlichia ewingii infections in the United States, 2008–2012. Am J Trop Med Hyg. 2016;94:52–60. http://dx.doi.org/10.4269/ ajtmh.15-0540 References 16. Harris RM, Couturier BA, Sample SC, Coulter KS, Casey KK, 1. Nadolny RM, Feldman KA, Pagac B, Stromdahl EY, Rutz H, Schlaberg R. Expanded geographic distribution and clinical Wee S-B, et al. Review of the Mid-Atlantic Tick Summit III: A characteristics of Ehrlichia ewingii infections, United States. Emerg model for regional information sharing. Ticks Tick Borne Dis. Infect Dis. 2016;22:862–5. http://dx.doi.org/10.3201/eid2205.152009 2015;6:435–8. http://dx.doi.org/10.1016/j.ttbdis.2015.04.001 17. Dahlgren FS, Heitman KN, Behravesh CB. Undetermined human 2. Centers for Disease Control and Prevention. Lyme disease data and ehrlichiosis and anaplasmosis in the United States, 2008–2012: statistics 2005–2014. 2015 [cited 2015 Dec 4]. http://www.cdc.gov/ a catch-all for passive surveillance. Am J Trop Med Hyg. lyme/stats/index.html 2016;94:299–301. http://dx.doi.org/10.4269/ajtmh.15-0691 3. Childs JE, Paddock CD. The ascendancy of Amblyomma 18. Karpathy SE, Slater KS, Goldsmith CS, Nicholson WL, americanum as a vector of pathogens affecting humans in Paddock CD. Rickettsia amblyommatis sp. nov., a spotted fever the United States. Annu Rev Entomol. 2003;48:307–37. group Rickettsia associated with multiple species of Amblyomma http://dx.doi.org/10.1146/annurev.ento.48.091801.112728 ticks in North, Central and South America. Int J Syst Evol

944 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Ehrlichiosis and Lyme Disease, New Jersey, USA

Microbiol. 2016;66:5236–43. http://dx.doi.org/10.1099/ land-use gradient. Ticks Tick Borne Dis. 2013;4:352–8. ijsem.0.001502 http://dx.doi.org/10.1016/j.ttbdis.2013.01.001 19. Stromdahl EY, Jiang J, Vince M, Richards AL. Infrequency of 34. Armstrong PM, Brunet LR, Spielman A, Telford SR III. Risk of Rickettsia rickettsii in Dermacentor variabilis removed from Lyme disease: perceptions of residents of a Lone Star tick-infested humans, with comments on the role of other human-biting ticks community. Bull World Health Organ. 2001;79:916–25. associated with spotted fever group Rickettsiae in the United 35. Hook SA, Nelson CA, Mead PSUS. U.S. public’s experience with States. Vector Borne Zoonotic Dis. 2011;11:969–77. ticks and tick-borne diseases: Results from national HealthStyles http://dx.doi.org/10.1089/vbz.2010.0099 surveys. Ticks Tick Borne Dis. 2015;6:483–8. http://dx.doi.org/ 20. Collins BR, Anderson KH. Plant communities of New Jersey: 10.1016/j.ttbdis.2015.03.017 a study in landscape diversity. New Brunswick (NJ): Rutgers 36. Walker DH. Ehrlichia under our noses and no one notices. University Press; 1994. In: Peters CJ, Calisher CH, editors. Infectious diseases from 21. Schulze TL, Jordan RA, White JC, Roegner VE, Healy SP. nature: mechanisms of viral emergence and persistence. Vienna: Geographical distribution and prevalence of selected Borrelia, Springer-Verlag/Wien; 2004. p. 147–56. Ehrlichia, and Rickettsia infections in Amblyomma americanum 37. Olano JP, Masters E, Hogrefe W, Walker DH. Human (Acari: Ixodidae) in New Jersey. J Am Mosq Control Assoc. monocytotropic ehrlichiosis, Missouri. Emerg Infect Dis. 2003; 2011;27:236–44. http://dx.doi.org/10.2987/11-6111.1 9:1579–86. http://dx.doi.org/10.3201/eid0912.020733 22. New Jersey Department of Health. 2005–2014 New Jersey 38. Yevich SJ, Sánchez JL, DeFraites RF, Rives CC, Dawson JE, Reportable Communicable Disease Reports [cited 2016 Mar 29]. Uhaa IJ, et al. Seroepidemiology of infections due to spotted fever https://www.nj.gov/health/cd/reportable_disease_stats.shtml group rickettsiae and Ehrlichia species in military personnel 23. Rollend L, Fish D, Childs JE. Transovarial transmission of exposed in areas of the United States where such infections are Borrelia spirochetes by Ixodes scapularis: a summary of the endemic. J Infect Dis. 1995;171:1266–73. http://dx.doi.org/ literature and recent observations. Ticks Tick Borne Dis. 10.1093/infdis/171.5.1266 2013;4:46–51. http://dx.doi.org/10.1016/j.ttbdis.2012.06.008 39. Standaert SM, Dawson JE, SchaffnerW, Childs JE, Biggie KL, 24. Long SW, Zhang X, Zhang J, Ruble RP, Teel P, Yu XJ. Singleton J Jr, et al. Ehrlichiosis in a golf-oriented retirement Evaluation of transovarial transmission and transmissibility of community. N Engl J Med. 1995;333:420–5. http://dx.doi.org/ Ehrlichia chaffeensis (Rickettsiales: Anaplasmataceae) in 10.1056/NEJM199508173330704 Amblyomma americanum (Acari: Ixodidae). J Med Entomol. 40. Fritz CL, Kjemtrup AM, Conrad PA, Flores GR, Campbell GL, 2003;40:1000–4. http://dx.doi.org/10.1603/0022-2585-40.6.1000 Schriefer ME, et al. Seroepidemiology of emerging tickborne 25. Loftis AD, Massung RF, Levin ML. Quantitative real-time infectious diseases in a Northern California community. J Infect PCR assay for detection of Ehrlichia chaffeensis. Dis. 1997;175:1432–9. http://dx.doi.org/10.1086/516476 J Clin Microbiol. 2003;41:3870–2. http://dx.doi.org/10.1128/ 41. Marshall GS, Jacobs RF, Schutze GE, Paxton H, Buckingham SC, JCM.41.8.3870-3872.2003 DeVincenzo JP, et al.; Tick-Borne Infections in Children Study 26. Killmaster LF, Loftis AD, Zemtsova GE, Levin ML. Detection Group. Ehrlichia chaffeensis seroprevalence among children in of bacterial agents in Amblyomma americanum (Acari: Ixodidae) the southeast and south-central regions of the United States. Arch from Georgia, USA, and the use of a multiplex assay to Pediatr Adolesc Med. 2002;156:166–70. http://dx.doi.org/10.1001/ differentiateEhrlichia chaffeensis and Ehrlichia ewingii. J Med archpedi.156.2.166 Entomol. 2014;51:868–72. http://dx.doi.org/10.1603/ME13225 42. Huhn GD, Sejvar JJ, Montgomery SP, Dworkin MS. West Nile 27. Unver A, Rikihisa Y, Stich RW, Ohashi N, Felek S. The omp-1 virus in the United States: an update on an emerging infectious major outer membrane multigene family of Ehrlichia chaffeensis disease. Am Fam Physician. 2003;68:653–60. is differentially expressed in canine and tick hosts. 43. Centers for Disease Control and Prevention. Ehrlichiosis: statistics Infect Immun. 2002;70:4701–4. http://dx.doi.org/10.1128/ and epidemiology, 1994–2010. 2015 [cited 2016 Jan 12]. IAI.70.8.4701-4704.2002 http://www.cdc.gov/ehrlichiosis/stats 28. Jaworski DC, Bowen CJ, Wasala NB. A white-tailed deer/lone 44. Ginsberg HS, Ewing CP, O’Connell AF Jr, Bosler EM, Daley JG, star tick model for studying transmission of Ehrlichia chaffeensis. Sayre MW. Increased population densities of Amblyomma Vector Borne Zoonotic Dis. 2013;13:193–5. http://dx.doi.org/ americanum (Acari: Ixodidae) on Long Island, New York. 10.1089/vbz.2011.0868 J Parasitol. 1991;77:493–5. http://dx.doi.org/10.2307/3283144 29. Starkey LA, Barrett AW, Chandrashekar R, Stillman BA, 45. Regan J, Matthias J, Green-Murphy A, Stanek D, Bertholf M, Tyrrell P, Thatcher B, et al. Development of antibodies to and PCR Pritt BS, et al. A confirmed Ehrlichia ewingii infection likely detection of Ehrlichia spp. in dogs following natural tick exposure. acquired through platelet transfusion. Clin Infect Dis. 2013; Vet Microbiol. 2014;173:379–84. http://dx.doi.org/10.1016/ 56:e105–7. http://dx.doi.org/10.1093/cid/cit177 j.vetmic.2014.08.006 46. Allen MB, Pritt BS, Sloan LM, Paddock CD, Musham CK, 30. Varela-Stokes AS. Transmission of Ehrlichia chaffeensis from Ramos JM, et al. First reported case of Ehrlichia ewingii lone star ticks (Amblyomma americanum) to white-tailed deer involving human bone marrow. J Clin Microbiol. 2014;52:4102–4. (Odocoileus virginianus). J Wildl Dis. 2007;43:376–81. http://dx.doi.org/10.1128/JCM.01670-14 http://dx.doi.org/10.7589/0090-3558-43.3.376 47. Centers for Disease Control and Prevention. Diagnosis and 31. Schulze TL, Jordan RA. The role of publicly owned properties in management of tickborne rickettsial diseases: Rocky Mountain the transmission of Lyme disease in central New Jersey. Journal of spotted fever, ehrlichioses, and anaplasmosis—United States: a prac- Spirochetal and Tick-borne Diseases. 1996;3:124–9. tical guide for physicians and other health-care and public health pro- 32. Yabsley MJ. Natural history of Ehrlichia chaffeensis: vertebrate fessionals. MMWR Morb Mortal Wkly Rep. 2006; 55(RR-4):1–29. hosts and tick vectors from the United States and evidence for endemic transmission in other countries. Vet Parasitol. 2010; Address for correspondence: Andrea Egizi, Rutgers University, 167:136–48. http://dx.doi.org/10.1016/j.vetpar.2009.09.015 33. Bayles BR, Evans G, Allan BF. Knowledge and prevention 180 Jones Ave, New Brunswick, NJ 08901, USA; email: of tick-borne diseases vary across an urban-to-rural human [email protected]

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 945 Distribution and Quantitative Estimates of Variant Creutzfeldt-Jakob Disease Prions in Tissues of Clinical and Asymptomatic Patients Jean Y. Douet, Caroline Lacroux, Naima Aron, Mark W. Head, Séverine Lugan, Cécile Tillier, Alvina Huor, Hervé Cassard, Mark Arnold, Vincent Beringue, James W. Ironside, Olivier Andréoletti

In the United-Kingdom, ≈1 of 2,000 persons could be infect- In 1996, a new form of CJD, termed variant CJD ed with variant Creutzfeldt-Jakob disease (vCJD). There- (vCJD) was identified in the United Kingdom. vCJD is fore, risk of transmission of vCJD by medical procedures believed to result from zoonotic transmission of the BSE remains a major concern for public health authorities. In this agent, probably as a consequence of dietary exposure to study, we used in vitro amplification of prions by protein mis- BSE-contaminated meat products (2,3). The total num- folding cyclic amplification (PMCA) to estimate distribution ber of clinical cases of vCJD thus far identified is lim- and level of the vCJD agent in 21 tissues from 4 patients who died of clinical vCJD and from 1 asymptomatic person ited (227 patients worldwide at the time of writing this with vCJD. PMCA identified major levels of vCJD prions in a article). However, the estimated prevalence of asymp- range of tissues, including liver, salivary gland, kidney, lung, tomatic vCJD in populations exposed to the BSE agent and bone marrow. Bioassays confirmed that the quantita- is uncertain (4). tive estimate of levels of vCJD prion accumulation provided In the United Kingdom, 32,441 appendix samples by PMCA are indicative of vCJD infectivity levels in tissues. collected during surgery from patients born during 1941– Findings provide critical data for the design of measures to 1985 have been tested for abnormal prion protein accu- minimize risk for iatrogenic transmission of vCJD. mulation by using immunohistochemical analysis. This study reported a vCJD prevalence estimate of 1/2,000 in rion diseases, or transmissible spongiform encepha- persons in these age cohorts (95% CI 1/3,500–1/1,250) Plopathies (TSEs), are fatal neurodegenerative disor- (5). No comparable data are available concerning the ders that occur naturally in sheep (scrapie), cattle (bo- prevalence of asymptomatic vCJD in other countries, vine spongiform encephalopathy [BSE]), and humans although BSE exposure is known to have occurred in (Creutzfeldt-Jakob disease [CJD]). A key event in the several countries in continental Europe, as judged by pathogenesis of TSEs is the conversion of the normal cel- cases of vCJD that are not attributable to exposure in the lular prion protein (PrPC, encoded by the PRNP gene) into United Kingdom (http://www.cjd.ed.ac.uk/documents/ an abnormal disease-associated isoform (PrPSc) in tissues worldfigs.pdf). of infected animals. PrPC is completely degraded after Over the past 2 decades, several studies have reported controlled digestion with proteinase K in the presence of on the distribution of the vCJD agent in tissues of infect- nondenaturing detergents. In contrast, PrPSc is N termi- ed patients (6–8). Most of these studies did not detect the nally truncated under the same conditions, resulting in a vCJD agent outside the nervous system (central, peripheral, proteinase K–resistant prion (PrPres) (1). and autonomic) and lymphoid tissues. However, the sensitivity of detection techniques for PrPres used in these inves- Author affiliations: Institut National de la Recherche Agronomique, tigations was limited. Toulouse, France (J.Y. Douet, C. Lacroux, N. Aron, S. Lugan, Protein misfolding cyclic amplification (PMCA) C. Tillier, A. Huor, H. Cassard, O. Andréoletti); University of is believed to mimic prion replication in vitro, but in an Edinburgh, Edinburgh, Scotland, UK (M.W. Head, J.W. Ironside); accelerated form, which enables amplification of minute Animal and Plant Health Agency, Loughborough, UK (M. Arnold); amounts of PrPSc and prion infectivity (9). In PMCA, a Institut National de la Recherche Agronomique, Jouy-en-Josas, PrPC-containing substrate is combined with a seed that France (V. Beringue) might contain otherwise undetectable amounts of PrPSc. DOI: http://dx.doi.org/10.3201/eid2306.161734 After repeated cycles of incubation and sonication, the

946 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Variant Creutzfeldt-Jakob Disease Prions

Sc amount of PrP increases to levels at which they can be Table 1. Endpoint titration of reference sample from a patient detected by using conventional biochemical techniques. with vCJD in tgBov mice expressing bovine prion protein* Recently, our group and others have shown that PMCA Transmission in tgBov mice No. positive Mean  SD, can detect endogenous vCJD agent in patient biologic flu- Dilution mice/no. tested incubation, d ids such as urine and blood (10,11). Undiluted 6/6 249  2 In this study, we evaluated the relative sensitivity of 101 6/6 283  15  PMCA versus that of bioassay in mice for detection of 10 2 6/6 316  21 103 6/6 342  10 the vCJD agent. We estimated by using PMCA the level  10 4 6/6 453  66 of vCJD prions in 21 tissues collected from 4 patients 105 2/6 479, 495† who died of symptomatic vCJD and from a patient with 106 1/6 502† asymptomatic vCJD. We also determined whether vCJD 107 0/6 >700 *A 10% (wt/vol) homogenate was prepared by using frontal cortex from a prion levels, as estimated by using PMCA, were consis- clinically affected patient with vCJD. Groups of 6 tgBov mice were tent with infectious titers, as estimated by bioassay with inoculated intracerebrally with 20 L of serial 10-fold dilutions of this homogenate. Mice were considered positive when abnormal prion protein transgenic mice. deposition was detected in the brain. vCJD, variant Creutzfeldt-Jakob disease. Methods †Dilutions at which <50% of mice were positive.

Ethics Statements vCJD and Control Patients All animal experiments were performed in compliance We investigated tissues from 4 clinical vCJD case-pa- with institutional and French national guidelines and in tients (vCJD-1–vCJD-4) and 1 asymptomatic person with accordance with the European Community Council Direc- vCJD who had received a transfusion of packed eryth- tive 86/609/EEC. Animal experiments that were part of this rocytes from a donor who subsequently died from vCJD study (national registration no. 01734.01) were approved (12). Tissues from 2 non–vCJD–affected patients were by the local ethic committee of the Ecole Nationale Vétéri- used as controls. For case-patients who provided appro- naire de Toulouse (Toulouse, France). Mouse inoculations priate consent, the entire PRNP gene coding sequence were performed under anesthesia with isofulorane. Mice was established to exclude pathogenic mutations in this that displayed clinical signs of disease were anesthetized gene (13,14). with isofluorane before being humanely killed by inhala- Mouse Bioassays tion of CO2. Human samples were obtained from the United King- Bioassays were performed by using mice expressing bo- dom National CJD Research and Surveillance Unit Brain vine PrP (tgBov-tg110) as described (15,16). These mice and Tissue Bank, which is part of the Medical Research were observed daily and their neurologic status was as- Council Edinburgh Brain Bank (Edinburgh, Scotland, sessed weekly. When clinically progressive TSE symptoms UK). Tissue samples were pseudo-anonymized by using were evident, or at the end of their lifespan, the animals a Brain Bank reference number. All case-patients in the were euthanized. Survival time was expressed as the mean United Kingdom provided informed consent. Use of sam- ± SD days postinoculation of mice positive for PrPres. For ples in this study was approved by the East of Scotland mice that showed no clinical signs, they were humanely Research Ethics Service for the Edinburgh Brain Bank killed at the end of their natural lifespan (600–800 days). (16/ES/0084). In these instances, incubation periods are reported as >600

Table 2. Endpoint titration of PMCA seeding activity in a reference brain sample from a patient with vCJD* Reference vCJD 10% brain homogenate dilution series, no. of positive PMCA reactions/no. reactions conducted Amplification round 102 103 104 105 106 107 108 109 1010 1 6/6 6/6 0/6 0/6 0/6 0/6 0/6 0/6 0/6 2 6/6 6/6 5/6 3/6 0/6 0/6 0/6 0/6 0/6 3 6/6 6/6 6/6 6/6 3/6 0/6 0/6 0/6 0/6 4 6/6 6/6 6/6 6/6 6/6 5/6 2/6 0/6 0/6 5 6/6 6/6 6/6 6/6 6/6 5/6 2/6 0/6 0/6 6 6/6 6/6 6/6 6/6 6/6 5/6 2/6 0/6 0/6 *A 10% (wt/vol) homogenate was prepared by using frontal cortex from a symptomatic patient with vCJD (same homogenate as in Table 1). Samples were serially diluted 10-fold (102–1010) before being used to seed PMCA reactions. Six individual replicates of each sample dilution were tested. PMCA substrate was prepared by using brains from transgenic mice overexpressing the ARQ variant of sheep prion protein. PMCAs were subjected to 6 rounds of amplification, each composed of 96 cycles (sonication for 10 s and incubation for 14.5 min at 39.5°C) in a Qsonica700 Sonicator (Qsonica LLC, Newtown, CT, USA). After each round, reaction products (1 volume) were mixed with fresh substrate (9 volumes) to seed the following round. Part of the same product was analyzed by Western blotting for abnormal PrPres by using Sha31 antibody epitope YEDRYYRE. Values are number of PrPres Western Blot–positive replicates corresponding to each round and each dilution. PMCA, protein misfolding cyclic amplification; PrPres, proteinase K–resistant prion; vCJD, variant Creutzfeldt-Jakob disease.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 947

RESEARCH days postinoculation, which corresponded to survival time The titer of prion infectivity in vCJD–affected patient observed for >3 of 6 mice. bone marrow samples was estimated by using the method of Arnold et al. (17). This method uses the probability of Estimation of Infectious Prion Titers survival (attack rate at each dilution) and the individual We estimated infectious titer in a reference 10% (wt/vol) mouse incubation periods at each dilution to estimate in- frontal cortex homogenate from a clinical vCJD patient by fectious load and is thus able to provide more accurate es- using endpoint titration (intracerebral route) in tgBov mice. timation of titer than using either attack rate or incubation Infectious titer (50% lethal dose/g intracerebral in tgBov period data alone. mice) was estimated by using the Spearman method. PMCA Reactions

A transgenic mouse line that expresses ovine A136R154Q171 PrP variant PrPC (tgShXI) was used to prepare the PMCA substrate as described (18,19). PMCA amplification was performed as described (11). Each PMCA experiment included a reference vCJD sample (10% brain homogenate) as a control for the amplification efficiency. Unseeded controls (1 unseeded control for 8 seeded reactions) were also included in each experiment. For each tested dilution of each sample, >4 replicates were tested in 2 independent experiments. For each sample, the highest dilution showing >50% of positive replicates (presence of detectable PrPres in the reaction as assessed by using Western blotting) was determined.

Detection of Abnormal PrP by Western Blotting and Paraffin-Embedded Tissue Blotting Extraction of proteinase K–resistant abnormal PrP and Western blotting were performed as described (11). Im- munodetection was performed by using 2 PrP-specific monoclonal antibodies, Sha31 (1 µg/mL) (20), and 12B2 (4 µg/mL) (21), which recognize amino acid sequences YEDRYYRE (145–152), and WGQGG (89–93), respectively. Paraffin-embedded tissue blotting was performed as described (22,23).

Results

Sensitivity of vCJD Agent Detection by Figure 1. Western blots of variant Creutzfeldt-Jakob disease PMCA and Bioassay (vJCD) proteinase K–resistant prions (PrPres) analyzed by protein To determine the relative sensitivity of PMCA, we reti- misfolding cyclic amplification (PMCA) in tissues of clinical and trated a reference sample (10% cerebral cortex homog- asymptomatic patients. PMCAs were seeded with a 10-fold enate from a vCJD-affected patient) that had previously serial dilutions of a reference vCJD brain homogenate (10% wt/ vol, 10-2–10-10 dilutions). This homogenate had been endpoint undergone endpoint titration (IC inoculation route; Table titrated by bioassay in bovine prion (PrP)–expressing mice (tgBov, 1) in bovine PrP–expressing mice (tgBov). Amplifica- intracerebral route, 107.7 50% lethal dose/g). PMCA substrate was tion of a 10-fold serial dilution of this sample (6 individ- prepared by using brains from transgenic mice overexpressing the ual replicates/dilution point) demonstrated that 4 PMCA ARQ variant of sheep prion protein. An unseeded (lane U) reaction rounds (24 hours/round, i.e., 96 h) were sufficient to reach was included as a specificity control. PMCAs were subjected to 6 rounds of amplification, each composed of 96 cycles (sonication the maximum sensitivity level of the assay. Additional for 10 s and incubation for 14.5 min at 39.5°C) in a Qsonica700 PMCA rounds did not improve the analytical sensitivity Sonicator (Qsonica LLC, Newtown, CT, USA). After each round, of the assay or the number of positive replicates (Table 2; reaction products (1 volume) were mixed with fresh substrate (9 Figure 1). On the basis of these results, we estimated by volumes) to seed the following round. Part of the same product res using the Spearman method that the seeding activity of was analyzed by Western blotting for abnormal PrP (Sha31 11 antibody epitope YEDRYYRE). A sheep scrapie sample (WB cont) the isolate was 10 50% seeding activity/per g. Bioassay was included as a control on each gel. WB, Western blot. endpoint titration data for the same sample in tgBov mice

948 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Variant Creutzfeldt-Jakob Disease Prions

Table 3. Characteristics of 5 patients with vCJD and 2 controls in study of distribution and quantitative estimates of variant prions in tissues* Patient Year of Age, y, at Disease PRNP gene PRNP gene identification no. Diagnosis Sex death death duration, mo codon 129 mutations vCJD-1 vCJD M 1999 33 18 MM None detected vCJD-2 vCJD F 2000 17 18 MM None detected vCJD-4 vCJD M 2000 26 10 MM None detected vCJD-3 vCJD M 2001 26 10 MM None detected vCJD-A Asymptomatic vCJD F 2004 82 NA MV None detected NC-1 No CJD (tumor, infarction, ischemia) F 2005 85 NA MM No consent for sequencing NC-2 No CJD (Alzheimer’s disease, F 2010 80 NA MM None detected infarction, ischemia) *NA, not applicable; vCJD, variant Creutzfeldt-Jakob disease.

7.7 showed an infectious titer of 10 LD50/g. When we took subjected to 4 rounds of PMCA. Amplification products into account the 4-fold lower amount of material used to from each round were tested for PrPres using by Western seed the PMCA reaction compared with material used in blotting (Table 4; Figure 2). mouse inoculations, we found that the PMCA protocol We found that none of the reactions seeded with tis- used was 465 times more sensitive than the bioassay of sue homogenates from non–CJD controls were positive for tgBov mice for detection of vCJD prions. PrPres (Table 4). In contrast, PMCA reactions seeded with tissues from the 4 symptomatic vCJD patients were posi- PMCA for Control and vCJD Patients tive for PrPres (Table 4; Figure 2). As expected, among test- We complied basic demographic data for vCJD and con- ed tissues, brain homogenates (temporal cortex) showed trol patients (Table 3). A 10-fold dilution series of 10% the highest seeding activity (highest PrPres-positive dilu- homogenates from the vCJD–affected and non–vCJD–af- tion 10−8). All lymphoid organs tested also showed seed- fected control patients was prepared, and this series was ing activity, but the highest PMCA-positive dilution varied

Table 4. Protein misfolding cyclic amplification reactions seeded with tissue homogenate from vCJD and control patients* Preclinical vCJD Non–vCJD controls, Clinical vCJD patients, Met129/Met129 patient, Met129/Val129 Met129/Met129 Tissue vCJD-1 vCJD-2 vCJD-3 vCJD-4 vCJD-A NC-1 NC-2 Frontal cortex 108 108 108 108 – – – Pituitary gland NA NA NA NA 102 – – Trigeminal ganglia NA NA NA NA – – – Dorsal root ganglia NA NA NA NA – – – Cervical lymph node 105 104 104 103 104 NA NA Tonsil 103 104 106 103 103 NA – Appendix 104 104 104 103 102 – – Distal Ileum 103 105 105 102 103 – – Spleen 104 104 105 104 103 – – Thymus NA 103 102 102 102 NA NA Lung 102 102 – – 103 – – Heart 102 102 – – – – – Liver 104 102 102 104 102 – – Kidney 102 103 – 103 – – – Salivary gland 104 103 102 103 102 – – Pancreas 102 – 102 104 – – – Thyroid 102 – 102 102 – – – Adrenal gland 103 103 103 104 – – – Bone marrow 104 105 103 104 – – – Skeletal muscle 104 102 – NA – – – Testis – NA – 103 NA NA NA Ovary NA 104 NA NA NA NA NA *PMCA reactions were seeded with 10-fold serial dilutions of 10% tissues homogenates (102–109) that had been collected postmortem from 4 symptomatic vCJD patients (vCJD-1–vCJD-4) or an asymptomatic vCJD-infected person (vCJD-A). At least 4 replicates of each sample dilution were tested in 2 independent PMCA experiments. Prions from patients vCJD-1–vCJD-4 were homozygous for methionine at codon 129 of the PRNP gene. Prion from patient vCJD-A was heterozygous (methionine/valine) at codon 129 of the PRNP gene. PMCA substrate was prepared by using brains from transgenic mice overexpressing the ARQ variant of sheep prion protein. Reactions seeded with tissues from 2 non–vCJD-infected control patients (NC-1 and NC-2) were included as negative controls. PMCAs were subjected to 4 rounds of amplification, each composed of 96 cycles (sonication for 10 s and incubation for 14.5 min at 39.5°C) in a Qsonica700 Sonicator (Qsonica LLC, Newtown, CT, USA). PMCA reactions were analyzed by Western blotting for proteinase K–resistant PrP by using Sha31 antibody epitope YEDRYYRE. Values are the highest dilution that resulted in a positive Western blot result in >50% of the tested replicates after 4 PMCA amplification rounds. NA, not applicable; PMCA, protein misfolding cyclic amplification; PrP, prion protein; vCJD, variant Creutzfeldt-Jakob disease; –, negative.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 949

RESEARCH

Figure 2. Protein misfolding cyclic amplification (PMCA) of peripheral tissues from patients with variant Creutzfeldt-Jakob disease (vCJD). PMCA reactions were seeded with a 10-fold dilution series of vCJD tissue homogenates (10-2–10−9) obtained postmortem from CJD-infected patients. At least 4 replicates of each sample dilution were tested in 2 independent PMCA experiments. Patients vCJD-1 (A), vCJD-2 (B), vCJD-3 (C), and vCJD-4 (D) died of clinical vCJD. These 4 patients were infected with prions containing methionine at codon 129 of the PRNP gene (homozygote). Patient vCJD-A (E) died during an asymptomatic or preclinical stage of the disease. This patient was infected with a prion containing methionine/valine at codon 129 of the PRNP gene (heterozygote). PMCA substrate was prepared by using brains from transgenic mice overexpressing the ARQ variant of sheep prion protein. Unseeded reactions and a reaction seeded with tissues from 2 non–vCJD-infected control patients (NC-1 and NC-2; Table 3) were included as specificity controls. PMCAs were subjected to 4 rounds of amplification, each composed of 96 cycles (sonication for 10 s and incubation for 14.5 min at 39.5°C) in a Qsonica700 Sonicator (Qsonica LLC, Newtown, CT, USA). After each round, reaction products (1 vol) were mixed with fresh substrate (9 vol) to seed the following round. Part of the same product was analyzed by Western blotting for abnormal proteinase K–resistant prions (PrPres) (Sha31 antibody epitope YEDRYYRE). For each round, the highest dilution showing a positive Western blot result in at least half of the replicates tested is indicated. Circles, round 1; ∇ round 2; ∆, round 3; squares, round 4. Neg, negative. according to the organs tested from 10-2 (thymus) to 10−6 was 103–106-fold lower than that for the brain. PrPres (distal ileum and tonsil). Moreover, for a given lymphoid was also detected by PMCA reactions seeded with heart, organ, <102-fold differences was observed in seeding activ- liver, kidney, skeletal muscle, several endocrine/exo- ity, depending on the patient and sample tested. These data crine glands (pancreas, thyroid), and gonads, from some, indicate that for symptomatic vCJD patients, lymphoid or- but not all, of the 4 clinical vCJD patients. Positive tis- gans contain 102–106-fold less prion seeding activity than sues contained a level of vCJD seeding activity that was the same amount of brain tissue (Table 4). equivalent to those observed in distal ileum (i.e., 103–106- Salivary gland, adrenal gland, liver, and bone mar- fold lower than for the brain). Irrespective of the tissue row from the 4 symptomatic vCJD patients showed posi- used to seed the PMCA reactions, the PrPres Western blot tive reactions by PMCA (Figures 2, 3). Using the highest profile for positive reactions was indistinguishable from dilution to show a positive reaction as a measure of seed- that observed in reactions seeded with the vCJD brain ing activity, we found that the vCJD agent in these tissues control (Figure 3).

950 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Variant Creutzfeldt-Jakob Disease Prions

Analysis of Tissues from an Asymptomatic vCJD-Infected Person Prion seeding activity was not detected in the brain (temporal cortex) of the asymptomatic vCJD–affected patient, who was infected with a PRNP gene codon 129 heterozygote (Met/Val129) prion (12) (Table 4; Figure 2). PMCA reactions seeded with dorsal root ganglia or trigeminal ganglia homogenates from this patient showed negative results. However, seeding activity was detected in the pituitary gland (highest PrPres-positive dilution 10-2). In addition, as for the symptomatic vCJD patient, PMCA amplification readily detected vCJD prions in all lymphoid organs tested from this asymptomatic person. On the basis of PMCA results, the vCJD agent load in lymphoid organs in this asymptomatic patient infected with the PRNP gene codon 129 Met/Val129 prion was similar to those for patients infected with Met/Met129 prions during the clinical stage of disease. In addition to findings for lymphoid organs, prion seeding activity was detectable in certain peripheral tissues (salivary gland, lung, and liver) from this patient (Tables 4; Figures 2, 3). Certain tissues, such as bone marrow or adrenal gland, that contained a substantial prion seeding activity in the clinically affected patients showed negative results. Again, the PrPres Western blot profile for positive reactions was indistinguishable from that observed for reactions seeded with the vCJD brain control. vCJD Infectivity in Bone Marrow To test whether PMCA seeding activity in peripheral tissues from vCJD patients correlated with infectivity, we inoculated bone marrow samples from the symptomatic patient into tgBov mice. Clinical TSE was observed in mice that were inoculated with each of the 4 bone marrow samples. The PrPres Western blot profile and the PrPres distribution pattern, as assessed by paraffin-embedded tissue blotting for brain of the bone marrow–inoculated mice, were identical to those observed in tgBov mice inoculated with the vCJD brain control sample (Figure 4). Figure 3. Western blots of proteinase K–resistant prions (PrPres) in Data obtained for mice inoculated with bone marrow PMCA reactions seeded with peripheral tissues. PMCA reactions -2 −9 samples were also used to estimate prion infectivity levels were seeded with a 10-fold dilution series (10 –10 ) of variant Creutzfeldt-Jakob disease (vCJD) tissue homogenates that had in these samples. For this purpose, we applied the method been collected postmortem from vCJD patients during the clinical of Arnold et al. (17). This method combines the probabil- stage (symptomatic vCJD patient 1–vCJD patient 4 [P1–P4]) or at ity of survival (attack rate) and the individual mouse incu- an asymptomatic or preclinical stage of the disease (vCJD asymp) bation period to provide an estimation of infectious titers. (Table 2). Reactions seeded with tissues from 2 non-vCJD patients We used data corresponding to endpoint titration in tgBov (Table 2) were used as controls, and an unseeded (lane U) reaction was included as a specificity control. Reactions were then subjected mice for reference vCJD sample (frontal cortex from a to 4 amplification rounds, each composed of 96 cycles (sonication clinical vCJD patient) (Table 1) to derive the relationship for 10 s and incubation for 14.5 min at 39.5°C) in a Qsonica700 between prion titer of inoculum and the probability of in- Sonicator (Qsonica LLC, Newtown, CT, USA). PMCA reactions fection and length of the incubation period (Figure 5). We were analyzed by using Western blotting for abnormal PrPres (Sha31 found that bone marrow samples had an infectious titer antibody epitope YEDRYYRE). A sheep scrapie sample and a vCJD reference isolate were used as controls. For the 7 tissues tested, the that ranged from 102.3 LD /g through 104.7 LD /g in tgBov 50 50 dilution of tissue homogenates used to seed the PMCA reactions is mice (Table 5). indicated below the immunoblots. Cont, control; WB, Western blot.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 951 RESEARCH

Figure 4. Detection of proteinase K–resistant prions (PrPres) by using Western blotting and paraffin-embedded tissue (PET) blotting of brains of transgenic mice expressing bovine PrP (tgBov). A) PrPres WB of a vCJD sample (frontal cortex), tgBov mice (brain) inoculated with the same vCJD reference isolate, bone marrow samples from vCJD- affected patients (vCJD 1–vCJD-4 [P1–P4]; Table 2), and a non– vCJD control (NC-1; Table 2). A scrapie isolate (WB cont) and a noninoculated tgBov brain (vCJD brain) homogenate were included as controls. PrPres immunodetection was performed by using Sha31 monoclonal antibody (epitope

145YEDRYYRE152) and 12B2 epitope

(epitope 89WGQGG93). B) PET blotting of PrPres distribution in coronal section (thalamus level) of tgBov mice inoculated with a reference vCJD isolate (10% brain homogenate) or bone marrow (10% tissue homogenate) from 2 vCJD patients (vCJD-1 and vCJD- 3; Table 2) at the clinical stage of disease. Immunodetection of PrPres was performed by using Sha31 monoclonal antibody (epitope

145YEDRYYRE152). Cont, control; NI, not inoculated; WB, Western blot. Scale bar indicates 120 µm.

These values are consistent with a 103–105 lower in- Natural scrapie and experimental BSE in sheep are 2 fectivity load in bone marrow samples than in the refer- models of orally transmitted prion diseases (24,25). In both ence vCJD brain sample. Consistent with the PMCA re- diseases, the agent accumulates in the lymphoreticular sys- sults (Table 4), we found that prion load in bone marrow tem and the enteric nervous system during the early pre- samples (highest PrPres–positive dilution [10-3–10−5]) was clinical phase of the incubation period. Moreover, an early also 103–105-fold lower than for the reference vCJD iso- and persistent prionemia is observed in asymptomatic in- late (highest PrPres–positive dilution [10−8]). These results fected animals (26,27). These features were also observed strongly support the idea that PMCA seeding activity pro- in vCJD in humans and in view of the likely origin of vCJD vides a reliable estimate of the prion load in tissues from (oral exposure to BSE agent), these similarities have led to vCJD-infected patients. a consensus that BSE and scrapie in sheep and vCJD in human have a common pathogenesis (28). Discussion Although vCJD prions in a variety tissues, such as Most previous studies with tissue from vCJD patients have bone marrow, kidney, salivary gland, skeletal muscle, failed to identify consistent accumulation of the vCJD pancreas, liver, or heart, might be surprising, each of agent outside the nervous and lymphoreticular systems. these tissue has already been demonstrated to accumulate However, data obtained in this study clearly demonstrate prion infectivity or abnormal prion protein in TSE-infect- the presence of vCJD prions in a wide and unexpected va- ed sheep (29–33). Because low levels of infectivity have riety of peripheral tissues. been reported in blood fractions from a vCJD-affected

952 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Variant Creutzfeldt-Jakob Disease Prions

trigeminal ganglia tissue from this patient are consistent with a lack of central nervous system involvement at the time of death. However, PMCA seeding activity in the pituitary gland was surprising in this instance. The presence of abnormal prion protein accumulation in the pituitary gland and other circumventricular organs before deposition of PrPres in surrounding brain has been reported in TSE-infected sheep (38). However, this phe- nomenon in animals does not represent the main route for neuroinvasion and is a probable consequence of hematogenous dissemination of the TSE agent through the fenes- trated capillary system of the circumventricular organs, which is substantially more permeable than the other capillaries in the brain (blood–brain barrier). Therefore, this finding might be a consequence of the hematogenous route of secondary vCJD in this person (by transfusion Figure 5. Dose–response relationship for A) incubation period and of packed erythrocytes from a vCJD-infected donor), in B) probability of infection of bovine PrP–expressing mice. Data contrast to the oral route of infection in primary clinical were derived from an endpoint titration of 10% (wt/vol) frontal vCJD cases (12). cortex homogenate from a patient with variant Creutzfeldt-Jakob vCJD prions were detected in certain peripheral tis- disease. This homogenate was inoculated into tgBov mice (20 µL by intracerebral [ic] route; Table 1). This procedure was used to sues from the patients infected with a prion containing the establish a model that estimates infectious titer in a homogenate PRNP gene codon 129 Met/Val. Although distribution of on the basis of incubation period and the probability of infection vCJD seeding activity in lymphoreticular tissues was simi- in inoculated mice. Model plots are indicated by solid lines. +, lar to that observed for symptomatic vCJD patients, sev- observed value. LD , 50% lethal dose. 50 eral tissues that were positive in clinically affected patients were negative in this heterozygous asymptomatic person. patient, such widespread tissue positivity might be de- These findings suggest that involvement of some peripheral rived from residual blood, rather than from the solid tis- tissues might occur at a later stage in the incubation pe- sue in these samples (16). However, this proposal seems riod than others, or that they could involve recirculation of unlikely because in whole blood PMCA amplification in- the agent from the central nervous system (i.e., centrifugal hibitors preclude detection of endogenous vCJD agent by spread in a late state). However, we cannot discount the this method (11,34–36). possibility that that these differences in tissue distribution The patient in our study who was infected with a prion are caused by the hematogenous route of infection in this containing PRNP gene codon 129 Met/Val is 1 of only 2 person (as opposed to the probable oral route in patients identified vCJD agent–infected persons known to have died with clinical vCJD) or the difference between the PRNP of other causes before onset clinical symptoms of vCJD, gene codon 129 genotype of the asymptomatic vCJD–af- and the only person who provided consent to sample au- fected person (PRNP gene codon 129 Met/Val) and persons topsy tissues for research. For this patient, all previous with clinical vCJD (PRNP gene codon 129 Met/Met). investigations did not detect abnormal prion protein or in- Irrespective of the actual explanation for these differ- fectivity in the brain (12,37). The negative PMCA results ences, the presence of vCJD agent in peripheral tissues of we obtained for cerebral cortex, dorsal root ganglia, and patients during preclinical and clinical stage of the disease

Table 5. Bone marrow sample bioassay in bovine PrP–expressing mice (tgBov) for 4 patients with vCJD* Transmission in tgBov mice

Sample No. positive/no. inoculated mice Mean  SD incubation, d Mean infectious titer, LD50/g (95% CI) vCJD-1 5/5 458  37 103.1 (102.6–103.5) vCJD-2 6/6 373  35 104.7 (104.3–05.2) vCJD-3 4/6 504  10 102.3 (101.8–102.7) vCJD-4 6/6 447  91 104.0 (103.4–04.5) PBS control 0/6 >600 NA *A 10% wt/vol bone marrow homogenate prepared from 4 symptomatic vCJD patients (Table 3) was inoculated intracerebrally into 6 tgBov mice (20 µL/mouse). One mouse (inoculated with homogenate from patient vCJD-1) died within the first few days after intercerebral inoculation. Mice were euthanized when they showed clinical signs of prion infection or after 600-d postinoculation. Mice were considered prion infected when abnormal PrP deposition was detected in brain. Infectious prion titers were estimated by using the method of Arnold et al. (17). The method uses the probability of survival (attack rate at each dilution) and the individual mouse incubation periods at each dilution to estimate infectious load. Infectious titers are given as estimated values. LD50, 50% lethal dose; NA, not applicable; PBS, phosphate-buffered saline; PrP; prion protein; vCJD, variant Creutzfeldt-Jakob disease.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 953

RESEARCH indicates the potential for iatrogenic transmission of this previous studies (11,16) and studies by Bougard et al. fatal neurologic condition by surgical procedures. Further- (35) and Concha-Marambio et al. (36). The relationship more, this finding shows that, for certain peripheral tissues, shown here between PrPres amplification by PMCA and a level of infectivity equivalent to an end stage titer (and detection of infectivity by bioassay indicates that PMCA attendant risk) is reached at a preclinical stage. seeding activity is a good surrogate marker of infectivity Several hundred cases of iatrogenic CJD have been re- and could provide a sound basis for a vCJD blood test for ported worldwide. These cases appear to result from trans- use with blood or tissue donors. mission of sporadic CJD, and most cases have occurred in This study was supported in part by the Department of Health recipients of human dura mater grafts or after administra- Policy Research Programme and the Scottish Government. The tion of human growth hormone extracted from cadaveric National CJD Research and Surveillance Unit is supported by pituitaries (39). Although in sporadic CJD the distribution the Policy Research Program of the Department of Health and of the agent is largely restricted to the nervous system (cen- the Scottish Government (DH121/5061). The Edinburgh Brain tral and peripheral), the wide distribution of the vCJD agent Bank is supported by the Medical Research Council (MRC grant in the asymptomatic infected patient we report might serve G0900580). The Unité Mixte de Recherche 1225, Ecole to increase the range of medical procedures, including den- Nationale Vétérinaire de Toulouse was supported by the tistry, organ transplant, and surgery involving nondispos- European Union FEDER/INTERREG (EFA282/13 able equipment, that might result in iatrogenic transmission TRANSPRION), the Institut National de la Recherche of vCJD (40–43). Agronomique Institut Carnot en Santé Animale, and an Agence Nevertheless, >20 years after identification of the first Nationale Recherche grant (Unmasking Blood Prions; vCJD patients, only 5 cases that are a probable consequence ANR-15-CE18-0028). of iatrogenic vCJD transmission are known, all in the Unit- ed Kingdom and associated with blood and blood products. Dr. Douet is a research scientist and assistant lecturer in These cases were caused by transfusion of non–leukocyte- ophthalmology at the National Veterinary School of Toulouse, depleted erythrocyte concentrates or by treatment involv- Toulouse, France. His primary research interests are the ing large amounts of pooled plasma from the United King- pathogenesis of the prion disease with special emphasis on the dom that were known to include donations from persons iatrogenic risk of transmission. who later showed development of vCJD (12,44–46). None of the 220 other vCJD cases identified world- References wide have been linked to any other medical or dental proce- 1. McKinley MP, Bolton DC, Prusiner SB. A protease-resistant dure. Whereas this fact is reassuring, it would be unwise to protein is a structural component of the scrapie prion. Cell. 1983;35:57–62. http://dx.doi.org/10.1016/0092-8674(83)90207-6 disregard the threat that vCJD still poses for public health. 2. Bruce ME, Will RG, Ironside JW, McConnell I, Drummond D, Despite the relatively low number (n = 178) of vCJD clini- Suttie A, et al. Transmissions to mice indicate that ‘new variant’ cal cases observed in the United Kingdom, the most recent CJD is caused by the BSE agent. Nature. 1997;389:498–501. epidemiologic studies indicate that ≈1 of 2,000 persons in http://dx.doi.org/10.1038/39057 3. Collinge J, Sidle KC, Meads J, Ironside J, Hill AF. Molecular the United Kingdom could be infected with the vCJD agent analysis of prion strain variation and the aetiology of ‘new (as indicated by the presence of abnormal prion protein de- variant’ CJD. Nature. 1996;383:685–90. http://dx.doi.org/ tected by immunohistochemical analysis of lymphoid fol- 10.1038/383685a0 licles in the appendix). Each asymptomatic vCJD-infected 4. Garske T, Ghani AC. Uncertainty in the tail of the variant Creutzfeldt-Jakob disease epidemic in the UK. PLoS One. person represents a potential source of secondary infection. 2010;5:e15626. http://dx.doi.org/10.1371/journal.pone.0015626 The data in our report offer an opportunity for refining mea- 5. Gill ON, Spencer Y, Richard-Loendt A, Kelly C, Dabaghian R, sures that were implemented in many countries to limit the Boyes L, et al. Prevalent abnormal prion protein in human risk for vCJD iatrogenic transmission. The apparent con- appendixes after bovine spongiform encephalopathy epizootic: large scale survey. BMJ. 2013;347:f5675. http://dx.doi.org/10.1136/ cordance between PMCA biochemical and infectivity bio- bmj.f5675 assay data, and the higher analytical sensitivity of PMCA, 6. Wadsworth JD, Joiner S, Hill AF, Campbell TA, Desbruslais M, suggest that future research need not rely exclusively on Luthert PJ, et al. Tissue distribution of protease resistant prion time-consuming and costly animal bioassay. protein in variant Creutzfeldt-Jakob disease using a highly sensitive immunoblotting assay. Lancet. 2001;358:171–80. Our results indicate the need for vCJD screening as- http://dx.doi.org/10.1016/S0140-6736(01)05403-4 says. After more than a decade of effort, several vCJD 7. Haïk S, Faucheux BA, Sazdovitch V, Privat N, Kemeny JL, blood detection tests have reached a stage in their devel- Perret-Liaudet A, et al. The sympathetic nervous system is involved opment that could enable their evaluation as screening in variant Creutzfeldt-Jakob disease. Nat Med. 2003;9:1121–3. http://dx.doi.org/10.1038/nm922 or confirmatory assays (11,47,48). In particular, there is 8. Head MW, Ritchie D, Smith N, McLoughlin V, Nailon W, now a strong case for use of PMCA in a highly sensi- Samad S, et al. Peripheral tissue involvement in sporadic, tive and specific blood test for vCJD, as indicated by our iatrogenic, and variant Creutzfeldt-Jakob disease: an

954 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Variant Creutzfeldt-Jakob Disease Prions

immunohistochemical, quantitative, and biochemical study. experimental oral transmission. J Gen Virol. 2001;82:2319–26. Am J Pathol. 2004;164:143–53. http://dx.doi.org/10.1016/ http://dx.doi.org/10.1099/0022-1317-82-10-2319 S0002-9440(10)63105-7 26. Lacroux C, Vilette D, Fernández-Borges N, Litaise C, Lugan S, 9. Saborio GP, Permanne B, Soto C. Sensitive detection of pathological Morel N, et al. Prionemia and leukocyte-platelet–associated prion protein by cyclic amplification of protein misfolding. Nature. infectivity in sheep transmissible spongiform encephalopathy 2001;411:810–3. http://dx.doi.org/10.1038/35081095 models. J Virol. 2012;86:2056–66. http://dx.doi.org/10.1128/ 10. Moda F, Gambetti P, Notari S, Concha-Marambio L, Catania M, JVI.06532-11 Park KW, et al. Prions in the urine of patients with variant 27. Houston F, Foster JD, Chong A, Hunter N, Bostock CJ. Creutzfeldt-Jakob disease. N Engl J Med. 2014;371:530–9. Transmission of BSE by blood transfusion in sheep. Lancet. 2000; http://dx.doi.org/10.1056/NEJMoa1404401 356:999–1000. http://dx.doi.org/10.1016/S0140-6736(00)02719-7 11. Lacroux C, Comoy E, Moudjou M, Perret-Liaudet A, Lugan S, 28. Hilton DA. Pathogenesis and prevalence of variant Creutzfeldt- Litaise C, et al. Preclinical detection of variant CJD and BSE Jakob disease. J Pathol. 2006;208:134–41. http://dx.doi.org/ prions in blood. PLoS Pathog. 2014;10:e1004202. http://dx.doi.org/ 10.1002/path.1880 10.1371/journal.ppat.1004202 29. Tamgüney G, Richt JA, Hamir AN, Greenlee JJ, Miller MW, 12. Peden AH, Head MW, Ritchie DL, Bell JE, Ironside JW. Wolfe LL, et al. Salivary prions in sheep and deer. Prion. Preclinical vCJD after blood transfusion in a PRNP codon 129 2012;6:52–61. http://dx.doi.org/10.4161/pri.6.1.16984 heterozygous patient. Lancet. 2004;364:527–9. http://dx.doi.org/ 30. Andréoletti O, Simon S, Lacroux C, Morel N, Tabouret G, 10.1016/S0140-6736(04)16811-6 Chabert A, et al. PrPSc accumulation in myocytes from sheep 13. Uro-Coste E, Cassard H, Simon S, Lugan S, Bilheude JM, incubating natural scrapie. Nat Med. 2004;10:591–3. Perret-Liaudet A, et al. Beyond PrPres type 1/type 2 dichotomy in http://dx.doi.org/10.1038/nm1055 Creutzfeldt-Jakob disease. PLoS Pathog. 2008;4:e1000029. 31. Chianini F, Cosseddu GM, Steele P, Hamilton S, Hawthorn J, http://dx.doi.org/10.1371/journal.ppat.1000029 Síso S, et al. Correlation between infectivity and disease associated 14. Moreno CR, Moazami-Goudarzi K, Laurent P, Cazeau G, prion protein in the nervous system and selected edible tissues of Andreoletti O, Chadi S, et al. Which PrP haplotypes in a French sheep naturally affected scrapie sheep. PLoS One. 2015;10:e0122785. population are the most susceptible to atypical scrapie? Arch Virol. http://dx.doi.org/10.1371/journal.pone.0122785 2007;152:1229–32. http://dx.doi.org/10.1007/s00705-007-0956-7 32. Hadlow WJ, Kennedy RC, Race RE. Natural infection of Suffolk 15. Castilla J, Gutiérrez Adán A, Brun A, Pintado B, Ramírez MA, sheep with scrapie virus. J Infect Dis. 1982;146:657–64. Parra B, et al. Early detection of PrPres in BSE-infected bovine http://dx.doi.org/10.1093/infdis/146.5.657 PrP transgenic mice. Arch Virol. 2003;148:677–91. 33. Garza MC, Monzón M, Marín B, Badiola JJ, Monleón E. http://dx.doi.org/10.1007/s00705-002-0958-4 Distribution of peripheral PrP(Sc) in sheep with naturally acquired 16. Douet JY, Zafar S, Perret-Liaudet A, Lacroux C, Lugan S, Aron N, scrapie. PLoS One. 2014;9:e97768. http://dx.doi.org/10.1371/ et al. Detection of infectivity in blood of persons with variant journal.pone.0097768 and sporadic Creutzfeldt-Jakob disease. Emerg Infect Dis. 2014; 34. Castilla J, Saá P, Soto C. Detection of prions in blood. Nat Med. 20:114–7. http://dx.doi.org/10.3201/eid2001.130353 2005;11:982–5. 17. Arnold ME, Hawkins SA, Green R, Dexter I, Wells GA. Pathogenesis 35. Bougard D, Brandel JP, Bélondrade M, Béringue V, Segarra C, of experimental bovine spongiform encephalopathy (BSE): Fleury H, et al. Detection of prions in the plasma of estimation of tissue infectivity according to incubation period. presymptomatic and symptomatic patients with variant Vet Res. 2009;40:8. http://dx.doi.org/10.1051/vetres:2008046 Creutzfeldt-Jakob disease. Sci Transl Med. 2016;8:370ra182. 18. Groschup MH, Buschmann A. Rodent models for prion diseases. http://dx.doi.org/10.1126/scitranslmed.aag1257 Vet Res. 2008;39:32. http://dx.doi.org/10.1051/vetres:2008008 36. Concha-Marambio L, Pritzkow S, Moda F, Tagliavini F, 19. Douet JY, Lacroux C, Corbière F, Litaise C, Simmons H, Lugan S, Ironside JW, Schulz PE, et al. Detection of prions in blood from et al. PrP expression level and sensitivity to prion infection. J Virol. patients with variant Creutzfeldt-Jakob disease. Sci Transl Med. 2014;88:5870–2. http://dx.doi.org/10.1128/JVI.00369-14 2016;8:370ra183. http://dx.doi.org/10.1126/scitranslmed.aaf6188 20. Féraudet C, Morel N, Simon S, Volland H, Frobert Y, Créminon C, 37. Bishop MT, Diack AB, Ritchie DL, Ironside JW, Will RG, et al. Screening of 145 anti-PrP monoclonal antibodies for their Manson JC. Prion infectivity in the spleen of a PRNP capacity to inhibit PrPSc replication in infected cells. J Biol Chem. heterozygous individual with subclinical variant Creutzfeldt-Jakob 2005;280:11247–58. http://dx.doi.org/10.1074/jbc.M407006200 disease. Brain. 2013;136:1139–45. http://dx.doi.org/10.1093/brain/ 21. Langeveld JP, Jacobs JG, Erkens JH, Bossers A, van Zijderveld FG, awt032 van Keulen LJ. Rapid and discriminatory diagnosis of scrapie and 38. Sisó S, González L, Jeffrey M. Neuroinvasion in prion diseases: BSE in retro-pharyngeal lymph nodes of sheep. BMC Vet Res. the roles of ascending neural infection and blood dissemination. 2006;2:19. http://dx.doi.org/10.1186/1746-6148-2-19 Interdiscip Perspect Infect Dis. 2010;2010:747892. 22. Cassard H, Torres JM, Lacroux C, Douet JY, Benestad SL, http://dx.doi.org/10.1155/2010/747892 Lantier F, et al. Evidence for zoonotic potential of ovine scrapie 39. Brown P, Brandel JP, Sato T, Nakamura Y, MacKenzie J, prions. Nat Commun. 2014;5:5821. http://dx.doi.org/10.1038/ Will RG, et al. Iatrogenic Creutzfeldt-Jakob disease, final ncomms6821 assessment. Emerg Infect Dis. 2012;18:901–7. http://dx.doi.org/ 23. Lacroux C, Corbière F, Tabouret G, Lugan S, Costes P, Mathey J, 10.3201/eid1806.120116 et al. Dynamics and genetics of PrPSc placental accumulation in 40. Flechsig E, Hegyi I, Enari M, Schwarz P, Collinge J, Weissmann C. sheep. J Gen Virol. 2007;88:1056–61. http://dx.doi.org/10.1099/ Transmission of scrapie by steel-surface-bound prions. Mol Med. vir.0.82218-0 2001;7:679–84. 24. Andréoletti O, Berthon P, Marc D, Sarradin P, Grosclaude J, 41. Taylor D. Inactivation of the BSE agent. C R Biol. 2002;325:75–6. van Keulen L, et al. Early accumulation of PrP(Sc) in gut-associated http://dx.doi.org/10.1016/S1631-0691(02)01386-0 lymphoid and nervous tissues of susceptible sheep from a 42. Molesworth A, Yates P, Hewitt PE, Mackenzie J, Ironside JW, Romanov flock with natural scrapie. J Gen Virol. 2000;81:3115–26. Galea G, et al. Investigation of variant Creutzfeldt-Jakob disease http://dx.doi.org/10.1099/0022-1317-81-12-3115 implicated organ or tissue transplantation in the United 25. Foster JD, Parnham DW, Hunter N, Bruce M. Distribution of the Kingdom. Transplantation. 2014;98:585–9. http://dx.doi. prion protein in sheep terminally affected with BSE following org/10.1097/TP.0000000000000105

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 955 RESEARCH

43. Jayanthi P, Thomas P, Bindhu P, Krishnapillai R. Prion diseases United Kingdom and in France. Transfusion. 2009;49:797–812. in humans: oral and dental implications. N Am J Med Sci. http://dx.doi.org/10.1111/j.1537-2995.2008.02044.x 2013;5:399–403. http://dx.doi.org/10.4103/1947-2714.115766 47. Edgeworth JA, Farmer M, Sicilia A, Tavares P, Beck J, Campbell T, 44. Peden A, McCardle L, Head MW, Love S, Ward HJ, Cousens SN, et al. Detection of prion infection in variant Creutzfeldt-Jakob et al. Variant CJD infection in the spleen of a neurologically disease: a blood-based assay. Lancet. 2011;377:487–93. asymptomatic UK adult patient with haemophilia. Haemophilia. http://dx.doi.org/10.1016/S0140-6736(10)62308-2 2010; 16:296–304. http://dx.doi.org/10.1111/j.1365-2516. 48. Jackson GS, Burk-Rafel J, Edgeworth JA, Sicilia A, Abdilahi S, 2009.02181.x Korteweg J, et al. A highly specific blood test for vCJD. Blood. 45. Llewelyn CA, Hewitt PE, Knight RS, Amar K, Cousens S, 2014;123:452–3. http://dx.doi.org/10.1182/blood-2013-11-539239 Mackenzie J, et al. Possible transmission of variant Creutzfeldt- Jakob disease by blood transfusion. Lancet. 2004;363:417–21. Address for correspondence: Olivier Andréoletti, Unité Mixte de http://dx.doi.org/10.1016/S0140-6736(04)15486-X Recherche 1225, Ecole Nationale Vétérinaire de Toulouse, 23 Chemin 46. Lefrère JJ, Hewitt P. From mad cows to sensible blood transfusion: the risk of prion transmission by labile blood components in the des Cappelles, Toulouse 31076, France; email: [email protected] etymologia Creutzfeldt-Jakob [croytsʹfelt-yakʺob] Disease Ronnie Henry, Frederick A. Murphy (L–R) Hans Gerhard Creutzfeldt (1885– 1964); Alfons Maria Jakob (1884–1931); Creutzfeldt–Jakob disease, cerebrum hematoxylin and eosin staining showing spongiform encephalophathy. Photographs reproduced from Foundations of Virology, 2012, courtesy Frederick A. Murphy. Micrograph courtesy of Sherif Zaki, Centers for Disease Control and Prevention, with permission.

n 1920, German neuropathologist Alfons Maria Jakob with genetic and iatrogenic forms. A zoonotic form, Idescribed a series of 6 patients with spasticity and variant CJD (vCJD), is caused by infection with a prion progressive dementia associated with neural degenera- derived from bovines and occurs predominantly in the tion. Shortly thereafter, in 1921, another German neuro- United Kingdom. pathologist, Hans Gerhardt Creutzfeldt, independently This issue of Emerging Infectious Diseases’ long- published a similar case. Jakob gave credit to Creutzfeldt running Etymologia series is dedicated to the memory of for describing the syndrome first, without realizing he had Richard T. Johnson, MD (1931–2015), the leading prion also uncovered the new syndrome. Walther Spielmeyer disease authority in the United States for many years and first used the term “Creutzfeldt-Jakob disease” (CJD) in great friend of CDC’s infectious disease programs, so 1922. CJD occurs worldwide as a rare, sporadic disease, many of which involve central nervous system disorders.

Sources 1. Brown P, Brandel JP, Sato T, Nakamura Y, MacKenzie J, Will RG, 3. Prusiner SB. Prions. Proc Natl Acad Sci U S A. 1998;95:13363– et al. Iatrogenic Creutzfeldt-Jakob disease, final assessment. Emerg 83. http://dx.doi.org/10.1073/pnas.95.23.13363 Infect Dis. 2012;18:901–7. http://dx.doi.org/10.3201/eid1806.120116 4. Richardson EP Jr, Masters CL. The nosology of Creutzfeldt-Jakob 2. Murphy FA. The foundations of virology: discoverers and disease and conditions related to the accumulation of PrPCJD discoveries, inventors and inventions, developers and technolo- in the nervous system. Brain Pathol. 1995;5:33–41. gies. West Conshohocken (PA): Infinity Publishing; 2012. http://dx.doi.org/10.1111/j.1750-3639.1995.tb00575.x Address for correspondence: Ronnie Henry, Centers for Disease Control and Prevention, 1600 Clifton Rd NE, Mailstop E03, Atlanta, GA 30329- 4027, USA; email: [email protected]

DOI: http://dx.doi.org/10.3201/eid2306.ET2306

956 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Outbreak-Related Disease Burden Associated with Consumption of Unpasteurized Cow’s Milk and Cheese, United States, 2009–2014 Solenne Costard, Luis Espejo, Huybert Groenendaal, Francisco J. Zagmutt

The growing popularity of unpasteurized milk in the United facilitated greater access of consumers to unpasteurized milk States raises public health concerns. We estimated outbreak- (e.g., through farm sales or cow share programs). The num- related illnesses and hospitalizations caused by the con- ber of states where the sale of unpasteurized milk is prohib- sumption of cow’s milk and cheese contaminated with Shiga ited decreased to 20 in 2011 from 29 in 2004 (8–10). This toxin–producing Escherichia coli, Salmonella spp., Listeria trend toward increased availability of unpasteurized dairy monocytogenes, and Campylobacter spp. using a model re- products raises public health concerns, especially because lying on publicly available outbreak data. In the United States, outbreaks associated with dairy consumption cause, on aver- raw milk consumers include children (2,4,6). age, 760 illnesses/year and 22 hospitalizations/year, mostly Our study aimed at estimating the outbreak-related from Salmonella spp. and Campylobacter spp. Unpasteurized disease burden associated with the consumption of fluid milk, consumed by only 3.2% of the population, and cheese, cow’s milk and cheese made from cow’s milk (herein also consumed by only 1.6% of the population, caused 96% of referred to as milk and cheese or dairy products) that are illnesses caused by contaminated dairy products. Unpas- unpasteurized and contaminated with Campylobacter spp., teurized dairy products thus cause 840 (95% CrI 611–1,158) Salmonella spp., Shiga toxin–producing Escherichia coli times more illnesses and 45 (95% CrI 34–59) times more (STEC), and Listeria monocytogenes. We also assessed hospitalizations than pasteurized products. As consumption how hypothetical increases in unpasteurized dairy con- of unpasteurized dairy products grows, illnesses will increase sumption would affect this outbreak-related disease burden. steadily; a doubling in the consumption of unpasteurized milk or cheese could increase outbreak-related illnesses by 96%. Methods

onsumer demand for organic and natural foods (i.e., Data Sources Cminimally processed foods) has been on the rise (1). We used outbreak data from the National Outbreak Report- However, in contrast to some perceptions (2), natural food ing System (NORS) (11) to estimate the incidence rates of products are not necessarily safer than conventional ones, as illnesses and hospitalizations. NORS is a web-based plat- evidenced by higher rates of foodborne illnesses associated form that stores data on all foodborne disease outbreaks with unpasteurized dairy products (3–6). Pasteurization has reported by local, state, and territorial health departments greatly reduced the number of foodborne illnesses attributed in the United States that have occurred since 2009. We to dairy products, and continuous efforts to reduce milk con- included all outbreaks that occurred during 2009–2014 tamination pre- and post-pasteurization are further decreas- in which the confirmed etiologic agents were any of the ing the disease burden (3). Yet, despite a decrease in dairy 4 pathogens of interest (Campylobacter spp., Salmo- consumption in the United States (7), recent studies (3,6) nella spp., STEC, and L. monocytogenes) and the impli- suggest that over the past 15 years the number of outbreaks cated food vehicle or contaminated ingredient was milk associated with unpasteurized dairy products has increased. or cheese (Figure 1). Outbreaks associated with multiple In parallel with this increase, an easing of regulations has products; processed dairy products other than milk and cheese (e.g., cream, butter, yogurt, and kefir); milk pro- Author affiliations: EpiX Analytics, Boulder, Colorado, USA duced by species other than cows; and cheese originat- (S. Costard, H. Groenendaal, F.J. Zagmutt); Consultant, ing from species other than cows were excluded from the St. Augustine, Florida, USA (L. Espejo) analysis (online Technical Appendix 1, https://wwwnc. cdc.gov/EID/article/23/6/15-1603-Techapp1.xlsx). DOI: https://dx.doi.org/10.3201/eid2306.151603

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 957 RESEARCH

Figure 1. Process for selecting US outbreaks associated with cow’s milk and cheese, 2009–2014. Laboratory- confirmed cases are cases with illness in which a specimen was collected and a laboratory was able to confirm the pathogen(s) or agent(s) causing illness. Hospitalizations are cases in which the patient was hospitalized as a result of becoming ill during the outbreak. NORS, National Outbreak Reporting System.

In addition, outbreaks with a suspected etiology status or hospitalization (λ) in the United States per serving of dairy associated with a dairy product with an unknown pasteuri- product using a conjugate gamma distribution (13). The num- zation status were excluded. ber of hospitalizations and laboratory-confirmed cases occur- The stochastic model (Figure 2) was developed to es- ring during the study period (2009–2014) that were caused timate the following: the incidence rates of illness and hos- by a given pathogen after consumption of milk or cheese of pitalization for pasteurized and unpasteurized dairy prod- a certain pasteurization status was obtained from the NORS ucts, the excess risk associated with unpasteurized milk and database. For laboratory-confirmed cases, this number was ad- cheese consumption, and the effect potential increases in justed for underreporting, under testing (only a proportion of consumption of unpasteurized dairy products would have suspected cases were sampled and tested), and underdiagnosis on the outbreak-related disease burden (online Techni- (based on diagnostic test sensitivity), in order to estimate ill- cal Appendix 2 Tables 1–5, https://wwwnc.cdc.gov/EID/ nesses for 2009–2014. These pathogen-specific factors were article/23/6/15-1603-Techapp2.pdf). Inputs (other than assumed to be independent of the product consumed and its the outbreak data) used in the stochastic model were de- pasteurization status, and constant for the years considered. rived from readily available sources of information (online The analysis did not include adjustment factors for potential Technical Appendix 2). Dairy consumption estimates were misclassification in terms of etiology or pasteurization status. derived from the Foodborne Active Surveillance Network These 2 outbreak characteristics were carefully reviewed, and (FoodNet) Population Survey (12). any outbreak for which the information could not be verified was excluded. It was thus assumed that etiology and pasteuri- Estimation of the Incidence of Outbreak-Related zation status misclassifications were negligible in this analysis. Illnesses and Hospitalizations Because NORS is a passive surveillance system, the We modeled the uncertainty of the pathogen-specific and inherent underreporting associated with it needed to be pasteurization status–specific incidence rates of illness and accounted for. We estimated an underreporting factor by

958 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Disease Burden and Unpasteurized Milk and Cheese

Figure 2. Stochastic model used to estimate the excess risk of outbreak-related illnesses and hospitalization due to unpasteurized dairy product consumption in the United States, 2009–2014. Model contains 3 main components: estimation of the incidence rates of illness and hospitalization for pasteurized and unpasteurized dairy products (elements in the boxes with solid lines), estimation of the excess risk associated with unpasteurized milk or cheese consumption (elements in box with dashed lines), and evaluation of the impact of hypothetical changes in consumption of unpasteurized dairy products (elements in boxes with dotted lines).

using FoodNet data, which is an active surveillance sys- States and that the food source attribution pertaining to tem assumed to include virtually all identified cases (on- the illnesses from confirmed and suspected outbreaks (16) line Technical Appendix 2). First, we extrapolated the total were equally relevant to laboratory-confirmed cases from number of laboratory-confirmed cases in the US -popula outbreaks of confirmed status only. We used the sensitivity tion during 2009–2013 using the incidence rates reported of the diagnostic tests as described in Scallan et al. (15) to by FoodNet and considering the proportions of the US estimate the proportion of false-negative, laboratory-con- population included in FoodNet surveillance sites (14). firmed cases from NORS (underdiagnosis factor). Finally, Second, we estimated the total number of outbreak-related we derived the under-testing factor by using the ratio of cases using the fraction of the US laboratory-confirmed laboratory-confirmed primary cases to the estimated total cases that were outbreak-related (15). Third, we extracted number of primary illnesses reported to NORS (17). the proportion of outbreak-related illnesses attributable to The annual number of servings of milk or cheese of a dairy (16). Fourth, we calculated the ratio of the number given pasteurization status was calculated as the product of of outbreak-related, laboratory-confirmed cases linked to the number of servings of milk or cheese per person for a dairy consumption derived from the previously described certain year, the resident population in the United States for calculations and the number of dairy-related, laboratory- that year (18) and the percentage of the population of dairy confirmed cases reported through NORS to use as the un- consumers that consume milk or cheese of a particular pas- derreporting factor in the analysis (online Technical Ap- teurization status. The annual per capita consumption of a pendix 2). When estimating the underreporting factor, we given dairy product (19) was divided by its average serving assumed that the FoodNet surveillance population and re- size (i.e., the amount of milk or cheese that is generally porting practices were representative of the entire United served) (7,20,21) to estimate the annual per capita number

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 959 RESEARCH of servings of milk and cheese. These totals were then We estimated the additional illnesses and the additional summed across the years of the study period. The per capita hospitalizations for each pathogen if a hypothetical increase consumption data (19) were assumed to include both pas- in consumption of unpasteurized milk or cheese occurred teurized and unpasteurized dairy products. Because unpas- using 1) the change in the proportion of the population teurized dairy products constitute a small percentage of the consuming unpasteurized milk or cheese, 2) the number of total consumption, this assumption (if inaccurate) would servings of milk or cheese for 2015, and 3) the risk differ- likely have only a small effect on results. We also hypoth- ence in illnesses per serving of dairy for that pathogen. We esized that the serving sizes (7,20,21) were the same for assumed that the overall proportion of the US population pasteurized and unpasteurized dairy products. consuming milk or cheese did not change; therefore, the The estimates of the proportion of dairy consumers increase in the proportion of the US population consum- that consume milk or cheese of a given pasteurization sta- ing unpasteurized milk or cheese corresponded to a shift tus were derived from the FoodNet Atlas of Exposure (12). of dairy consumers from pasteurized to unpasteurized. Six Answers from this FoodNet survey are provided as aggre- hypothetical scenarios were considered: 10%, 20%, 50%, gates per survey site, rather than per respondent. There- 100%, 200%, and 500% increases in the proportion of the fore, answers regarding milk and cheese consumption US population consuming unpasteurized milk or cheese. were treated as independent. In addition, we assumed that respondents who reported consumption of unpasteurized Scenario and Sensitivity Analyses milk or cheese did not consume pasteurized milk or cheese. We performed a sensitivity analysis to identify the param- Because the information to calculate the overall proportion eters that most influenced our estimates. The sensitivity of of the US population consuming any type of cheese was the estimates to the input parameter uncertainties was cal- unavailable, we assumed it to be equal to the proportion of culated by using conditional means as implemented in @ the population reporting consumption of any cheese sold as RISK 6.1.2 (Palisade Corporation, Ithaca, NY, USA). In or cut from solid blocks (i.e., the type of cheese consumed addition, we assessed the robustness of our sensitivity anal- most commonly). We further assumed the proportion of the ysis with a scenario analysis in which we calculated our US population consuming unpasteurized cheese to be equal estimates with different sets of outbreak data. For the main to the proportion reporting exposure to any cheese made analysis, the model was run on outbreaks of confirmed from unpasteurized milk in the previous 7 days. etiology and pasteurization status. In the scenario analysis, the model was then re-run with either of the 2 follow- Estimation of the Excess Risks Attributed to the ing sets of outbreaks added to the main data set: outbreaks Consumption of Unpasteurized Milk and Cheese of suspected etiology status (17) and outbreaks involving We estimated the additional risks for illness and hospital- dairy products of unspecified pasteurization status assumed ization for consumers of unpasteurized dairy products com- to be caused by pasteurized dairy products. pared with consumers of pasteurized ones. We calculated excess risk using 1) risk difference (RD), which measures Model Implementation the absolute difference in the observed risks for illness and The model was developed in Excel 2010 (Microsoft Corpo- hospitalization between consumers of unpasteurized dairy ration, Redmond, WA, USA) with the Monte-Carlo simu- products and consumers of pasteurized ones, and 2) inci- lation add-in @RISK 6.1.2. Results are expressed as means dence rate ratio (IRR), which provides a relative compari- and 95% credibility intervals (CrIs, a Bayesian equivalent son of the risks for illness and hospitalization between the to the confidence interval) or prediction intervals (PIs, 2 exposure groups (22). which provides uncertainty bounds for predictions), unless stated otherwise. Effects of Hypothetical Changes in Consumption of Unpasteurized Milk or Cheese Results We assessed the potential public health effects of hypothetical changes in unpasteurized milk consumption. We deter- Incidence Rates and Increased Risks Associated with mined the number of illnesses in 2015 in the United States the Consumption of Unpasteurized Milk and Cheese using the pathogen-specific rates of illnesses and hospital- We used a total of 87 outbreaks causing 750 laboratory- izations per serving of dairy product. The number of hospi- confirmed illnesses and 215 hospitalizations in this analysis talizations was calculated as pathogen-specific fractions of (Table 1). The incidence rates of STEC, Salmonella spp., these illnesses. The pathogen-specific probabilities of hos- and Campylobacter spp. illnesses and hospitalizations per 1 pitalization in cases of illness were assumed unconditional billion servings were higher for unpasteurized dairy product on the pasteurization status of the dairy product involved, consumers than for pasteurized dairy product consumers. but rather dependent on the severity of illness (23,24). Illnesses and hospitalizations caused by L. monocytogenes

960 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Disease Burden and Unpasteurized Milk and Cheese

Table 1. Dairy-related illnesses and hospitalizations from 87 outbreaks, National Outbreak Reporting System, United States, 2009– 2014* Outbreaks associated with milk and cheese consumption, N = 87† Pasteurized Unpasteurized Pathogen Outbreaks Illnesses Hospitalizations Outbreaks Illnesses Hospitalizations STEC 0 0 0 14‡ 99 42 Salmonella spp. 0 0 0 8§ 83 29 Listeria monocytogenes 10 100 87 1 1 1 Campylobacter spp. 1 2 0 53‡§ 465 56 Overall 11 102 87 76 648 128 *Illnesses and hospitalizations had confirmed etiologies and were associated with the consumption of milk or cheese of known pasteurization status. STEC, Shiga toxin–producing Escherichia coli. †Out of the 87 outbreaks, 10 outbreaks reported a total of 17 deaths, 16 of them were linked to L. monocytogenes and 1 to Campylobacter spp. ‡One outbreak (38 illnesses and 10 hospitalizations) had 3 cases with confirmed coinfection (STEC and Campylobacter spp.). These 3 cases were duplicated because they were assigned to each pathogen. §One outbreak (4 illnesses and 1 hospitalization) involved 2 pathogens: 3 Illnesses and 1 hospitalization were linked to Campylobacter spp. and 1 illness and 0 hospitalizations were linked to Salmonella spp. infections were more often attributed to the consumption of for unpasteurized dairy consumers than for pasteurized dairy pasteurized cheese than unpasteurized cheese (Table 2). As- consumers (mean IRR 45.1, 95% CrI 33.7–59.2), with an suming no change in the consumption of unpasteurized dairy, IRR of 0.5 (95% CrI 0–1.7) for L. monocytogenes. dairy products contaminated with STEC, Salmonella spp., L. monocytogenes, and Campylobacter spp. were predicted to Effects of Hypothetical Scenarios cause 761 (95% PI 598–994) outbreak-related illnesses and If the percentage of unpasteurized milk consumers in the 22 (PI 13–32) hospitalizations in 2015. Unpasteurized dairy United States were to increase to 3.8% and unpasteurized products caused 96% (PI 94%–98%) of these illnesses. cheese consumers to 1.9% (i.e., an increase of 20%), the We calculated the excess risk attributable to the con- number of illnesses per year would increase by an aver- sumption of unpasteurized milk and cheese (Table 2; Figure age of 19% and the number of hospitalizations by 21%. If 3). Because no reported illnesses were caused by Salmonella the percentages of unpasteurized milk and cheese consum- spp. and STEC during 2009–2014 and no hospitalizations ers were to double, the number of illnesses would increase were caused by Campylobacter spp., the corresponding in- by an average of 96%, and the number of hospitalizations cidence rates were extremely low (Table 2). Therefore, only would increase by 104%, resulting in an additional 733 RDs (and not IRRs) were reported for these pathogens. If (95% PI 571–966) illnesses/year and 22 (95% PI 13–32) all milk and cheese consumed were pasteurized, an average hospitalizations/year, which corresponds to a total of 1,493 of 732 (95% PI 570–966) illnesses and 21 (95% PI 12–32) (95% PI 1,180–1,955) illnesses/year (Figure 4), most hospitalizations would be prevented per year in the United caused by Salmonella spp. and Campylobacter spp. States. Of these prevented cases, 54% would be salmonellosis and 43% campylobacteriosis. The mean IRR of illnesses Scenario and Sensitivity Analyses was 838.8 (95% CrI 611.0–1,158.0) overall from all 4 patho- The following conditional means sensitivity analysis re- gens of interest (Figure 3), with 0.4 (95% CrI 0–1.2) from ports the change in the output mean if the input variable L. monocytogenes and 7,601 (95% CrI 3,711–15,346) from is set to its 5th and 95th percentiles while other inputs are Campylobacter spp. The rate of hospitalization was higher sampled at random. The rates of illnesses (λ) caused by the

Table 2. Incidence rates and risk differences for illness and hospitalization per 1 billion servings of milk or cheese, by pasteurization status and pathogen, United States, 2009–2014* Illnesses Hospitalizations Risk Risk Pathogen Unpasteurized Pasteurized difference† Unpasteurized Pasteurized difference† STEC 3.5 3.4 x 104 (3.1 x 3.5 0.9 3.4 x 104 (3.0 x 0.9 (2.7–4.5) 107 to 1.7 x 103) (2.7 to 4.5) (0.6 to 1.2) 107 to 1.7 x 103) (0.6 to 1.2) Salmonella spp. 49.1 3.4 x 104 (3.3 x 49.1 0.6 3.5 x 104 (3.4 x 0.6 (32.7–76.7) 107 to 1.7 x 103) (32.7 to 76.7) (0.4 to 0.9) 107 to 1.7 x 103) (0.4 to 0.9) Listeria 0.04 0.1 0.06 0.03 0.06 0.03 monocytogenes (0.003–0.100) (0.08 to 0.12) (0.11 to 0.02) (2.2 x 103 to 0.1) (0.05 to 0.07) (0.06 to 0.04) Campylobacter 39.0 5.8x103 (2.4 x 103 39.0 1.2 3.5 x 104 (3.5 x 1.2 spp. (30.8–48.3) to 1.1 x 102) (30.8 to 48.3) (0.9 to 1.5) 107 to 1.7 x 103) (0.9 to 1.5) Overall 91.7 0.11 91.6 2.7 6.1 x 102 (4.9 x 2.7 (71.8–120.9) (0.09 to 0.13) (71.7 to 120.8) (2.2 to 3.3) 102 to 7.5 x 102) (2.2 to 3.2) *Values are shown as mean incidence (95% credibility interval). STEC, Shiga toxin–producing Escherichia coli. †Excess risk is attributable to unpasteurized dairy.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 961

RESEARCH

pasteurized cheese, albeit only causing 1 additional outbreak-related illness per year on average. An easing of regulations has allowed greater access to unpasteurized milk in recent years (8–10), and this study shows that illnesses and hospitalizations will rise as consumption of unpasteurized dairy products increases. If such consumption were to double, the mean number of outbreak-related illnesses that occur every year would increase by 96%. Most unpasteurized dairy–related outbreaks are Figure 3. Forest plot showing, on a logarithmic scale, the excess caused by pathogen contamination at the dairy farm (ver- risk for outbreak-related illnesses and hospitalizations caused by consumption of pasteurized and unpasteurized milk and cheese, sus postpasteurization contamination for pasteurized prod- United States, 2009–2014. Markers indicate mean log IRR of ucts) (3); thus, one could assume that decreasing pathogen outbreak-related illnesses and hospitalizations caused by the prevalence in bulk milk tanks on raw milk farms would food pathogens Campylobacter spp., Listeria monocytogenes, help reduce illnesses. STEC has been found in 2.5% (95% Salmonella spp., and Shiga toxin–producing Escherichia coli CrI 0.1%–9.1%), Salmonella spp. in 4.6% (3.7%–5.6%), L. per 1 billion servings of unpasteurized milk or cheese relative to pasteurized products. Error bars indicate 95% credibility interval monocytogenes in 2.5% (0.1%–9.0%), and Campylobacter (CrI). Numbers above markers and bars are the IRR (not in spp. in 4.7% (2.8%–7.0%) of bulk milk tanks on US raw log scale) and 95% CrI. log (IRR) = 0 indicates no difference in milk farms (25–29). Given these low prevalences, strate- incidence rates between unpasteurized and pasteurized milk and gies for further reduction are limited and involve multiple cheese. IRR, incidence rate ratio. aspects of unpasteurized milk production (30). Boiling of milk before consumption seems to be a more realistic consumption of unpasteurized milk and cheese were most mitigation strategy, but this practice is unlikely to be imple- sensitive to the underreporting factors (γ) for Salmonella mented by unpasteurized dairy product advocates because spp. (mean range λ 34.9–72.5), Campylobacter spp. (mean it would affect the perceived benefits. range λ 33.1–45.3), and STEC (mean range λ 3.1–4.1), and This study focused on the outbreak-related illnesses, at a secondary level to the undertesting (ρ) and underdi- which is only a fraction of all dairy-related illnesses in agnosis (μ) factors (results not shown). The overall IRR the United States. Two studies have documented the frac- of illnesses was most sensitive to the underreporting factor tion of outbreak-related cases among FoodNet laboratory- for Salmonella spp. (mean range IRR 710.1–1,049.6). The confirmed cases (15,31); the fraction ranges from 0.5% for number of illnesses per year caused by the consumption of Campylobacter spp. to 19.0% for STEC according to Ebel milk or cheese was most sensitive to the rates of illnesses et al. (31). These data suggest that the number of sporad- caused by Salmonella spp. and Campylobacter spp., as the ic illnesses caused by contaminated dairy products in the main uncertainties apply to the incidence calculations for United States might be much larger than that for outbreak- all pathogens (results not shown). Including the 9 outbreaks related illnesses. However, because of the lack of infor- with a suspected-etiology status or the outbreak of unspeci- mation on the characteristics of sporadic illnesses (such fied pasteurization status (Figure 1) into the main analysis as food source attribution), we restricted the scope of this did not change the IRRs or the predicted number of illness- analysis to outbreak-related disease burden. es or hospitalizations per year (results not shown). Our analysis relied on outbreak data from NORS (11), which is a passive reporting system affected by underreport- Discussion ing. We used dairy-related outbreak cases from FoodNet Unpasteurized dairy products are responsible for almost (14–16) as a comparison to estimate underreporting; there- all of the 761 illnesses and 22 hospitalizations in the fore, any potential bias of this comparison was carried over United States that occur annually because of dairy-related to our estimation of outbreak-related illnesses. By extrapo- outbreaks caused by STEC, Salmonella spp., L. mono- lating incidence rates of cases from the FoodNet catchment cytogenes, and Campylobacter spp. More than 95% of areas to the overall United States, we assumed that the these illnesses are salmonellosis and campylobacteriosis. FoodNet surveillance population and reporting practices Consumers of unpasteurized milk and cheese are a small were representative of the entire United States. However, proportion of the US population (3.2% and 1.6%, respec- the FoodNet catchment population represents only 15% of tively), but compared with consumers of pasteurized dairy the US population from 10 nonrandom sites. Also, a recent products, they are 838.8 times more likely to experience study (31) suggested state-to-state variations in reporting an illness and 45.1 times more likely to be hospitalized. practices; these variations might be even greater between Illnesses caused by L. monocytogenes, however, were FoodNet and non-FoodNet states. This difference might found to be more often associated with the consumption of influence state-specific incidence rates or underreporting

962 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Disease Burden and Unpasteurized Milk and Cheese

or cheese only represented a fraction of their dairy consumption, the number of servings of unpasteurized dairy products could have been overestimated, and thus the risk for consumers of unpasteurized dairy products might have been underestimated. Also, the FoodNet population survey is based on a relatively small convenience sample and might therefore not be accurate. For example, the self-reported estimates of consumption of unpasteurized milk and cheese (3.2% and 1.6%, respectively) (12) might be underestimates or overestimates, potentially caused by consumers confus- ing the terms raw, organic, and natural (or other reasons). In addition, consumption might have changed since the 2007 FoodNet population survey (12), which might have resulted Figure 4. Number of dairy-related outbreak illnesses predicted per year in the United States if unpasteurized cow’s milk and cheese in an under- or overestimation of the risk from unpasteurized consumption increases 0%, 10%, 20%, 50%, 100%, 200%, and milk products. However, because the proportion of dairy 500%. Numbers in parentheses indicate percentage of total consumers using unpasteurized products remains small, and population consuming unpasteurized cow’s milk. The illnesses the IRRs are very large, this overestimation is likely limited, graphed are those from outbreaks associated with cow’s milk or and the trend for additional illnesses as unpasteurized dairy cheese contaminated with Shiga toxin–producing Escherichia coli, Salmonella spp., Listeria monocytogenes, and Campylobacter consumption grows remains valid. Similarly, estimates of spp. Markers indicate means; bars indicate 95% prediction the consumption of pasteurized cheese are underestimates: intervals. The consumption estimates were based on the year data available only provide estimates of the highest expo- 2015, and a 0% increase corresponds to the current proportion of sure to a single type of cheese, rather than to any type of the US population consuming unpasteurized dairy products. cheese (12), potentially resulting in a risk overestimation for consumers of pasteurized dairy products. This is a limita- ratios, as well as other characteristics of the reported cas- tion, notably for outbreaks linked to queso fresco and other es. For example, if a state reported the incriminated food Mexican-style soft cheeses. Despite these limitations, to the source as the food item (e.g., homemade yogurt), it would authors’ knowledge, this study is based on the best avail- not have been selected for inclusion in this analysis, but if able data and builds upon other well accepted risk attribution they reported the ingredient used for preparation (e.g., in methods (15,16,32). the case of homemade yogurt, fluid milk), it would have In conclusion, outbreaks linked to the consumption of been included in our analysis. However, the size and direc- cow’s milk and cheese were estimated to cause on average tion of such biases and uncertainties associated with these 761 illnesses and 22 hospitalizations per year in the United complex surveillance systems (NORS and FoodNet) are States. Unpasteurized products are consumed by a small difficult to quantify because of the paucity of data. percentage of the US dairy consumers but cause 95% of The rates of illnesses were most sensitive to the estimat- illnesses; the risk for illness was found to be >800 times ed underreporting factors, which were assumed to be associ- higher for consumers of unpasteurized milk or cheese than ated with the severity of symptoms (23,24) and other factors, for consumers of pasteurized dairy products. Therefore, such as state health department resources, and thus indepen- outbreak-related illnesses will increase steadily as unpas- dent of the pasteurization status. Also, because this analysis teurized dairy consumption grows, likely driven largely by only considered outbreaks involving milk and cheese (and no salmonellosis and campylobacteriosis. other dairy products), we are probably underestimating the incidence of illnesses and hospitalizations. However, milk Acknowledgments and cheese are the most commonly consumed dairy sourc- The authors thank collaborators E. Hovingh, D.R. Wolfgang, es and cause the most outbreaks (milk and cheese caused and H. Lysczek for their valuable input on milk production and 99% of dairy-related outbreaks reported to NORS during the consumption. We also thank the anonymous reviewers for their study period), so the underestimation is likely limited. None- helpful suggestions. theless, the overall comparison of risk between consumers of This work was supported in part by a United States Department pasteurized and unpasteurized products should remain valid. of Agriculture Special Research Grant (no. 2010-34163-21179) Estimates of the proportion of the population consum- and the Pennsylvania Agricultural Experiment Station. ing dairy products were derived from the FoodNet population survey (12). We assumed that respondents who reported Dr. Costard is an epidemiologist working as a senior consultant consumption of unpasteurized milk or cheese were not con- at EpiX Analytics in Boulder, Colorado. Her general research suming pasteurized dairy. However, if unpasteurized milk interests include risk analysis, simulation modeling, and

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 963 RESEARCH quantitative decision support tools; she has special interests in outbreak data, United States, 1998–2008. Emerg Infect Dis. health risk management strategies and food safety. 2013;19:407–15. http://dx.doi.org/10.3201/eid1903.111866 17. Centers for Disease Control and Prevention. NORS: National Outbreak Reporting System guidance. Atlanta (GA): Centers for Disease Control and Prevention; 2012 [cited 2016 Apr 29]. References https://www.cdc.gov/nors/pdf/NORS_Guidance_5213-508c.pdf 1. Food Marketing Institute. Natural and organic foods. Arlington 18. United States Census Bureau. Population estimates. 2016 (VA): Food Marketing Institute; 2008 [cited 2015 Sep 29]. [cited 2016 Apr 29]. https://www.census.gov/programs-surveys/ https://www.fda.gov/ohrms/dockets/dockets/06p0094/06p- popest.html 0094-cp00001-05-Tab-04-Food-Marketing-Institute-vol1.pdf 19. United States Department of Agriculture, Economic Research 2. KatafiaszAR, Bartlett P. Motivation for unpasteurized milk Service. Dairy data. 2014 [cited 2016 Apr 29]. http://www.ers.usda. consumption in Michigan, 2011. Food Prot Trends. 2012; gov/data-products/dairy-data.aspx#.UqJ_meJrfYh 32:124–8. 20. Sebastian RS, Goldman JD, Enns CW, LaComb RP. Fluid milk 3. Langer AJ, Ayers T, Grass J, Lynch M, Angulo FJ, Mahon BE. consumption in the United States: what we eat in America, Nonpasteurized dairy products, disease outbreaks, and state laws— NHANES 2005–2006. 2010 Sep [cited 2016 Sept 29]. United States, 1993–2006. Emerg Infect Dis. 2012;18:385–91. https://www.ars.usda.gov/ARSUserFiles/80400530/pdf/DBrief/3_ http://dx.doi.org/10.3201/eid1803.111370 milk_consumption_0506.pdf 4. Robinson TJ, Scheftel JM, Smith KE. Raw milk consumption 21. Innovation Center for US Dairy. Diary’s role in the diet: among patients with non–outbreak-related enteric infections, Cheese. 2013 [cited 2017 Apr 05]. http://www.usdairy.com/ Minnesota, USA, 2001–2010. Emerg Infect Dis. 2014;20:38–44. DairyResearchInstitute/NHanes/Pages/Cheese.aspx http://dx.doi.org/10.3201/eid2001.120920 22. Gordis L. Epidemiology. 2nd ed. Philadelphia: W.B. Saunders; 2000. 5. Claeys WL, Cardoen S, Daube G, De Block J, Dewettinck K, 23. Haas CN, Rose JB, Gerba CP. Quantitative microbial risk Dierick K, et al. Raw or heated cow milk consumption: assessment. New York: John Wiley & Sons; 1999. review of risks and benefits. Food Control. 2013;31:251–62. 24. Interagency Microbiological Risk Assessment Guideline work- http://dx.doi.org/10.1016/j.foodcont.2012.09.035 group. Microbial risk assessment guideline: pathogenic organisms 6. Mungai EA, Behravesh CB, Gould LH. Increased outbreaks with focus on food and water. United States Department of associated with nonpasteurized milk, United States, 2007–2012. Agriculture, Food Safety and Inspection Service; Emerg Infect Dis. 2015;21:119–22. http://dx.doi.org/10.3201/ Environmental Protection Agency; 2012 [cited 2015 Sept 29]. eid2101.140447 https://www.epa.gov/sites/production/files/2013-09/documents/ 7. Stewart H, Dong D, Carlson A. Why are Americans consuming less mra-guideline-final.pdf fluid milk? A look at generational differences in intake frequency, 25. Jayarao BM, Henning DR. Prevalence of foodborne pathogens in ERR-149. 2013 May [cited 2015 Sep 29]. https://www.ers.usda. bulk tank milk. J Dairy Sci. 2001;84:2157–62. http://dx.doi.org/ gov/webdocs/publications/err149/37651_err149.pdf?v=41423 10.3168/jds.S0022-0302(01)74661-9 8. National Association of State Departments of Agriculture. 26. Jayarao BM, Donaldson SC, Straley BA, Sawant AA, Hegde NV, Raw milk survey, November, 2004 [cited 2016 Apr 29]. Brown JL. A survey of foodborne pathogens in bulk tank milk and http://www.nasda.org/File.aspx?id=1582 raw milk consumption among farm families in Pennsylvania. 9. National Association of State Departments of Agriculture. NASDA J Dairy Sci. 2006;89:2451–8. http://dx.doi.org/10.3168/jds.S0022- releases raw milk survey. 2008 Apr 21 [cited 2016 Apr 29]. 0302(06)72318-9 https://view.officeapps.live.com/op/view.aspx?src=http%3A%2F%2F 27. Van Kessel JS, Karns JS, Gorski L, McCluskey BJ, Perdue ML. www.nasda.org%2FFile.aspx%3Fid%3D2149 Prevalence of salmonellae, Listeria monocytogenes, and fecal 10. National Association of State Departments of Agriculture. NASDA coliforms in bulk tank milk on US dairies. J Dairy Sci. 2004;87:2822– releases raw milk survey. 2011 Jul 19 [cited 2016 Apr 29]. 30. http://dx.doi.org/10.3168/jds.S0022-0302(04)73410-4 http://www.nasda.org/file.aspx?id=3916 28. Van Kessel JA, Karns JS, Lombard JE, Kopral CA. Prevalence 11. Centers for Disease Control and Prevention. National Outbreak of Salmonella enterica, Listeria monocytogenes, and Reporting System (NORS). Atlanta (GA): Centers for Disease Escherichia coli virulence factors in bulk tank milk and in-line Control and Prevention; 2009 [cited 2016 Apr 29]. filters from U.S. dairies. J Food Prot. 2011;74:759–68. https://www.cdc.gov/nors/index.html http://dx.doi.org/10.4315/0362-028X.JFP-10-423 12. Centers for Disease Control and Prevention. Foodborne Diseases 29. Jayarao BM, Hovingh. Raw milk–is it safe? In: American Dairy Active Surveillance Network (FoodNet) population survey atlas Science Association, American Society of Animal Science, of exposures, 2006-2007. Atlanta (GA): Department of Health and Canadian Society of Animal Science. 2014 Joint Annual Human Services, Centers for Disease Control and Prevention Meeting abstracts. J Anim Sci. 2014;92(E-Suppl 2):147; J Dairy [cited 2016 Sept 29]. https://www.cdc.gov/foodnet/surveys/ Sci. 2014;97(E-Suppl 1):147. FoodNetExposureAtlas0607_508.pdf 30. Hoenig DE. Raw milk production: guidelines for Maine licensed 13. Gelman A, Carlin JB, Stern HS, Dunson DB, Vehtari A, dealers. 2014 Sep 4 [cited 2015 Sep 29]. http://umaine.edu/ Rubin DB. Bayesian data analysis. 3rd ed. London: CRC Press, publications/1030e/ Taylor & Francis Group; 2013. 31. Ebel ED, Williams MS, Cole D, Travis CC, Klontz KC, Golden NJ, 14. Centers for Disease Control and Prevention. Foodborne Diseases et al. Comparing characteristics of sporadic and outbreak- Active Surveillance Network (FoodNet). Atlanta (GA): Centers associated foodborne illnesses, United States, 2004–2011. for Disease Control and Prevention. 2014 [cited 2016 Apr 29]. Emerg Infect Dis. 2016;22:1193–200. http://dx.doi.org/10.3201/ https://www.cdc.gov/foodnet/index.html eid2207.150833 15. Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV, Widdowson MA, 32. Williams MS, Ebel ED, Vose D. Framework for microbial food- Roy SL, et al. Foodborne illness acquired in the United States— safety risk assessments amenable to Bayesian modeling. Risk Anal. major pathogens. Emerg Infect Dis. 2011;17:7–15. 2011;31:548–65. http://dx.doi.org/10.1111/j.1539-6924.2010.01532.x http://dx.doi.org/10.3201/eid1701.P11101 16. Painter JA, Hoekstra RM, Ayers T, Tauxe RV, Braden CR, Angulo FJ, et al. Attribution of foodborne illnesses, Address for correspondence: Solenne Costard, EpiX Analytics, 1643 hospitalizations, and deaths to food commodities by using Spruce St, Boulder, CO 80302, USA; email: [email protected]

964 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 DISPATCHES

Sustainability of High-Level Isolation Capabilities among US Ebola Treatment Centers

Jocelyn J. Herstein, Paul D. Biddinger, outbreak revealed vulnerabilities within the US healthcare Shawn G. Gibbs, Aurora B. Le, Katelyn C. Jelden, and public health infrastructure to address HIDs. We aimed Angela L. Hewlett, John J. Lowe to identify barriers to maintenance of recently developed isolation and care capabilities, how those capabilities might To identify barriers to maintaining and applying capabili- be applied to outbreaks other than EVD, and further infra- ties of US high-level isolation units (HLIUs) used during the structure and resources HLIUs would add if additional Ebola virus disease outbreak, during 2016 we surveyed funding were available. HLIUs. HLIUs identified sustainability challenges and reported the highly infectious diseases they would treat. HLIUs expended substantial resources in development but must The Study strategize models of sustainability to maintain readiness. In early 2016, we sent a 70-question survey to the original 56 designated US HLIUs, including the 10 RESPTCs. The survey queried challenges and concerns about the maintenance uring the 2014–2016 West Africa Ebola virus disease of capabilities. Results were collected through Adobe Ac- D(EVD) outbreak, 56 hospitals in the United States were robat Pro (https://acrobat.adobe.com/us/en/acrobat/acrobat- designated by the Centers for Disease Control and Preven- pro.html) and analyzed by using descriptive statistics. The tion as Ebola treatment centers (ETCs). ETCs added national University of Nebraska Medical Center Institutional Review capacity to care for patients with highly infectious diseases Board declared the study exempt (#172–16X). (HIDs); that is, hazardous, easily transmissible, life-threat- Thirty-six (64%) hospitals responded. Of the 33 that ening illnesses with limited treatment options, such as EVD completed the full survey, 3 reported they no longer main- and severe acute respiratory syndrome coronavirus (1). ETCs tained their HLIU capabilities. The 2 that provided qualitative were equipped with the clinical care resources, specialized information about their decision to close reported needing infrastructure, and trained staff to safely manage and treat HLIU resources for other, more pressing areas and cited close a person suspected or confirmed to have EVD 2( ). After the proximity to at least 1 other HLIU as reasons for closing. initial designation, 1 ETC in each US Department of Health Nineteen (58%) hospitals reported using their HLIU and Human Services region was selected as a Regional Ebola for non-HID patients when not activated; the other 14 and Other Special Pathogens Treatment Center (RESPTC) (42%) use the unit exclusively for patients with HIDs or capable of managing HIDs for extended periods (3). for training (Table 1). When the 19 hospitals with adaptive In 2009, a consensus group of infectious disease ex- isolation units (i.e., units otherwise used for normal hospi- perts in Europe defined high-level isolation units (HLIUs) tal care) are activated, an average of 6.31 beds (median 6, as facilities providing optimal infection containment and range 2–12) must be taken offline when caring for 1 patient procedures specifically designed for HID care and released with an HID and an average of 6.97 beds (median 7.75, specifications for such units 1( ). A 2015 pilot survey of US range 2–12) for 2 patients. Ten (53%) HLIUs with adaptive HLIUs described the actions taken to establish high-level units stated preference for a unit dedicated to care for pa- isolation capabilities and identified the costs of those efforts tients with HIDs; however, when asked the estimated costs (4–6). The survey revealed that 45 of the US hospitals spent of developing a unit for 2 HID patients, estimates ranged a cumulative total of $53.9 million (nearly $1.2 million per from $1 million to $12 million. Perceived benefits of a ded- facility) to stand up their specialized isolation units (4). icated unit included minimizing disruption of other patients Because of the substantial expenses and operation- (4 hospitals), a constant state of readiness (3 hospitals), and al challenges of maintaining readiness, how HLIUs can an ability to train in the unit (4 hospitals). continue these efforts has been questioned (7). The EVD Our initial 2015 survey reported that hospitals designated Author affiliations: University of Nebraska Medical Center, as ETCs incurred an average per hospital of $1,197,993 (4). Omaha, Nebraska, USA (J.J. Herstein, K.C. Jelden, A.L. Hewlett, Since that time, 25 (76%) of those original facilities reported J.J. Lowe); Harvard Medical School, Boston, Massachusetts, receiving some degree of federal reimbursement, and 8 (24%) USA (P.D. Biddinger); Indiana University School of Public Health, had not received any reimbursement. A cumulative total of Bloomington, Indiana, USA (S.G. Gibbs, A.B. Le) $28,146,558 in federal funding (average $1,407,328, range $33,650–$6,000,000) was reported by the 20 (60%) report- DOI: https://dx.doi.org/10.3201/eid2306.170062 ing HLIUs. After we excluded federally funded RESPTCs

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 965 DISPATCHES

Table 1. Activation of HLIUs and management of PUIs, United States* Variable Facilities, no./N (%) Activation of HLIU HLIU can be activated 24/7 throughout the year† 32/33 (97) Standing protocol exists to contact team members 24/7 31/33 (94) Involve local/state public health officials in managing public concerns 32/33 (97) PUIs Plan to provide care for PUIs and persons with confirmed cases 32/33 (97) Staff used to care for PUI Use only HLIU staff to care for a PUI 28/32 (88) Use other staff before disease is confirmed 4/32 (13) Placement of PUI Place PUI exclusively in the HLIU while being assessed 14/32 (44) Place PUI in either HLIU or hospital ED 12/32 (38) Place PUI in ED until confirmed diagnosis 4/32 (13) Other‡ 2/32 (6) *ED, emergency department; HLIU, high-level isolation unit; PUI, patient under investigation. †Average time necessary to activate HLIU after notification of pending patient transfer is 4.58 h (median 4 h, range 1.24 h). ‡One facility sends a mobile response team to a PUI’s home for evaluation, and another plans to use a mobile treatment unit (i.e., tent) for PUI placement. and HLIUs that did not report initial investments in the pilot facility modifications and/or capabilities they would add if survey, the remaining 14 HLIUs reported a gap in reimburse- additional hypothetical funding were available (Table 2). ment of $9,113,072.50 (mean $650,933.75 per HLIU). Although 1 HLIU reported lacking specific protocols Conclusions or an ability to care for patients with an HID other than Developed during the height of the West Africa Ebola EVD, all other HLIUs (97%) reported being prepared to outbreak, most newly established US HLIUs invested care for patients with HIDs other than EVD (Figure 1). Our immense resources and effort into preparing for patients survey also queried HLIUs about the challenges they ex- with EVD. However, no formal network of HLIUs has perienced and challenges they foresee in maintaining the been established in the United States, except for the 10 capabilities and capacity needed for HID care (Figure 2). RESPTCs, and at least 3 former HLIUs no longer maintain Sustainability concerns was the most cited challenge in es- HLIU capabilities. Moreover, 14 HLIUs not designated as tablishing and maintaining a HLIU. HLIUs also detailed RESPTCs reported having spent $9.1 million more than they have been reimbursed to initially develop HLIU capabilities. As a result, these hospitals reported struggling to fund ongoing operations and sustain readiness. Although many facilities have created adaptable-use HLIUs because they lack the capital funds, space, or both to create a dedicated unit, such units have major disadvantages

Figure 2. Challenges to establishing an HLIU and to maintaining HID care reported by survey respondents, United States, 2016 (n = 32 HLIUs). Other challenges include external support, lack of dedicated unit space, competing priorities, staffing needs, and Figure 1. Diseases that 31 high-level isolation units (HLIUs) decreasing hospital capacities. HLID, high-level isolation unit; reported they would treat, United States, 2016. HID, highly infectious disease.

966 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Sustainability of US Ebola Treatment Centers

Table 2. Operational capabilities HLIUs reported they would add or construct if funding were available, United States* Funding amount Capability No. HLIUs $100,000 Additional training/drills (e.g., for other diseases, purchase of simulation equipment) 6 Broad supplies/equipment (e.g., beds, ventilators, family support technology/equipment) 4 Laboratory capability and capacity (e.g., reduced transport of materials, lab hood in unit, purchase 4 of new decontamination equipment) $500,000 On-site waste disposal 4 Expanded and updated patient rooms 3 Enhanced laboratory capabilities (e.g., additional laboratory tests, larger lab space) 3 Expanded isolation unit (e.g., increase capacity of negative-pressure rooms) 2 $1,000,000 Renovated/expanded isolation unit 4 Separate, permanent isolation unit 3 Expanded training (e.g., increased frequency) 2 *Individual HLIDs self-reported data through an electronically administered survey administered in 2016. HLIU, high-level isolation unit. because healthcare workers are unable to train in the unit, Public Health. Her research interests include highly infectious existing patients must be relocated when the unit is activated diseases, global high-level isolation units, infectious disease for an HID patient, and multiple additional rooms must be response and preparedness, and infection control training of taken off-line for the care of 1 patient with an HID 8( ). Thus, non–healthcare workers. more than half of US HLIUs that routinely care for non- HID patients would build an HID-dedicated unit if funds References were available. However, because future funding sources for 1. Bannister B, Puro V, Fusco FM, Heptonstall J, Ippolito G; EUNID non-RESPTCs are unclear, lessons on sustainability might Working Group. Framework for the design and operation of be learned from flexible-use HLIUs in Italy and the Nether- high-level isolation units: consensus of the European Network of lands, which offer levels of containment based on a patient’s Infectious Diseases. Lancet Infect Dis. 2009;9:45–56. http://dx.doi.org/10.1016/S1473-3099(08)70304-9 condition and offset costs by routine use 1( ). 2. Centers for Disease Control and Prevention. Interim guidance for Our study had several limitations. The data were self- U.S. hospital preparedness for patients under investigation (PUIs) reported and not validated by external sources. The current or with confirmed Ebola virus disease (EVD): a framework for a status of HLIUs that did not participate in the follow-up tiered approach [cited 2016 October 12]. http://www.cdc.gov/vhf/ ebola/healthcare-us/preparing/hospitals.html survey is unknown. A decrease in participation from the 3. US Department of Health and Human Services. HHS selects initial survey to the follow-up could also be due to the lon- regional Ebola treatment center for southwestern U.S. [cited 2016 ger, more detailed follow-up and could indicate the lack of Oct 15]. https://www.hhs.gov/about/news/2016/06/14/hhs-selects- attention to this area now that the EVD outbreak is over. regional-ebola-treatment-center-southwestern-us.html 4. Herstein JJ, Biddinger PD, Kraft CS, Saiman L, Gibbs SG, The study population was based solely on a list published Smith PW, et al. Initial costs of Ebola treatment centers in the by the Centers for Disease Control and Prevention (9) and United States. Emerg Infect Dis. 2016;22:350–2. http://dx.doi.org/ does not include data from other hospitals that similarly 10.3201/eid2202.151431 tried to strengthen their ability to treat HID patients. 5. Herstein JJ, Biddinger PD, Kraft CS, Saiman L, Gibbs SG, Le AB, et al. Current capabilities and capacity of Ebola treatment centers in In conclusion, a network of hospitals capable of the United States. Infect Control Hosp Epidemiol. 2016;37:313–8. treating patients with HIDs was rapidly constructed in re- http://dx.doi.org/10.1017/ice.2015.300 sponse to the recent EVD outbreak. However, without the 6. Jelden KC, Iwen PC, Herstein JJ, Biddinger PD, Kraft CS, impending threat of EVD or another HID on the immedi- Saiman L, et al. US Ebola treatment center clinical laboratory support. J Clin Microbiol. 2016;54:1031–5. http://dx.doi.org/ ate horizon, public attention on HID preparedness tends to 10.1128/JCM.02905-15 waver, and governments tend to prioritize and shift fund- 7. Dzau VJ, Sands P. Beyond the Ebola battle—winning the war ing elsewhere. Additional external funding sources re- against future epidemics. N Engl J Med. 2016;375:203–4. main generally uncertain for US HLIUs not designated as http://dx.doi.org/10.1056/NEJMp1605847 8. Schilling S, Fusco FM, De Iaco G, Bannister B, Maltezou HC, RESPTCs; therefore, these HLIUs must strategize meth- Carson G, et al.; European Network for Highly Infectious Diseases ods and models of sustainability if they are to maintain project members. Isolation facilities for highly infectious diseases capabilities and readiness. in Europe—a cross-sectional analysis in 16 countries. PLoS One. 2014;9:e100401. http://dx.doi.org/10.1371/journal.pone.0100401 Acknowledgment 9. Centers for Disease Control and Prevention. Current Ebola treatment centers [cited 2016 Oct 12]. https://www.cdc.gov/vhf/ We gratefully acknowledge Philip Smith for his expertise and ebola/healthcare-us/preparing/current-treatment-centers.html insight in developing the follow-up survey.

Ms. Herstein is a PhD student at the University of Nebraska Address for correspondence: Jocelyn J. Herstein, College of Medical Center and a graduate research assistant for the Public Health, University of Nebraska Medical Center, 984388 Nebraska Biocontainment Unit and Department of Environ- Nebraska Medical Center, Omaha, NE 68198, USA; email: mental, Occupational, and Agricultural Health in the College of [email protected]

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 967

DISPATCHES

Clinical and Molecular Characteristics of Human Rotavirus G8P[8] Outbreak Strain, Japan, 2014

Kenji Kondo, Takeshi Tsugawa, Mayumi Ono, and 2 clinics) in the cities of Sapporo, Tomakomai, Muro- Toshio Ohara, Shinsuke Fujibayashi, ran, and Urakawa (Figure 1). Yasuo Tahara, Noriaki Kubo, Shuji Nakata, For each fecal sample, we prepared a 10% fecal suspen- Yoshihito Higashidate, Yoshiki Fujii, sion, from which we extracted viral RNA. We performed Kazuhiko Katayama, Yuko Yoto, Hiroyuki Tsutsumi reverse transcription PCR on the RNA by using the Super- Script II Reverse Transcriptase (Invitrogen, Carlsbad, CA, During March–July 2014, rotavirus G8P[8] emerged as USA); PrimeSTAR GXL DNA polymerase (Takara, Shiga, the predominant cause of rotavirus gastroenteritis among Japan); and previously described primers (3,4). We used the children in Hokkaido Prefecture, Japan. Clinical character- BigDye Terminator v.3.1 Cycle Sequencing Reaction Kit istics were similar for infections caused by G8 and non-G8 strains. Sequence and phylogenetic analyses suggest the (Applied Biosystems, Foster City, CA, USA) to sequence strains were generated by multiple reassortment events be- PCR amplicons. For some of the rotavirus samples, next- tween DS-1–like P[8] strains and bovine strains from Asia. generation sequencing was performed at the National In- stitute of Infectious Diseases in Tokyo, Japan, as described previously (5). Sequences of the rotaviruses used in this otaviruses, the leading cause of acute gastroenteritis in study were submitted to the DDBJ under accession num- Rchildren worldwide, are classified into G and P geno- bers LC102884–LC103134 and LC105000–LC105532. types on the basis of 2 outer capsid proteins, viral protein We successfully determined G and P genotypes for (VP) 7 and VP4. A recently established extended rotavirus 148 of the 165 rotavirus samples by using the RotaC ro- genotyping system based on the sequence of all 11 genome tavirus genotyping tool (http://rotac.regatools.be/). The segments (1) grouped most human rotaviruses into 2 geno- most common genotype was G8P[8], which was identi- type constellations: Wa-like (G1/3/4/9-P[8]-I1-R1-C1-M1- fied in 58 samples (39.2%), followed by G1P[8] (25.7%), A1-N1-T1-E1-H1) and DS-1–like (G2-P[4]-I2-R2-C2- G9P[8] (20.3%), and G2P[4] (12.8%) (Table 1; on- M2-A2-N2-T2-E2-H2) strains. line Technical Appendix, https://wwwnc.cdc.gov/EID/ In industrialized countries, rotavirus genotype G8 in- article/23/6/16-0038-Techapp1.pdf). fection is common in bovines but rarely occurs in humans; We obtained clinical data for all 84 patients who however, the G8 strains are highly prevalent among hu- sought care at the hospital or clinic in Tomakomai. Demo- mans in some countries in Africa (2). We investigated the graphic and clinical characteristics (e.g., age, sex, history clinical and molecular features of G8P[8] rotavirus, which, of rotavirus vaccination, duration of fever, and duration we unexpectedly found to be the predominant genotype in and frequency of diarrhea and vomiting) were not substan- southwestern Hokkaido Prefecture, Japan, in 2014. tially different between 42 patients with G8P[8] rotavirus infection and 42 patients with non-G8P[8] rotavirus infec- The Study tion. The proportion of patients admitted to hospitals was During March–July 2014, we obtained rotavirus-positive also similar in the 2 groups (Table 1). fecal samples from 165 children in Hokkaido with acute We selected 15 G8P[8] strains for whole-genome gastroenteritis. The children were receiving care as inpa- analysis. All strains had the same genotype constellation, tients or outpatients at 1 of 6 medical facilities (4 hospitals G8-P[8]-I2-R2-C2-M2-A2-N2-T2-E2-H2, indicating a ge- Author affiliations: Sapporo Medical University School of Medicine, nomic backbone of the DS-1 genotype constellation. The Sapporo, Japan (K. Kondo, T. Tsugawa, M. Ono, Y. Yoto, genomes of these G8P[8] strains shared >99.6% nt identity H. Tsutsumi); Tomakomai City Hospital, Tomakomai, Japan with each other (Table 2; online Technical Appendix). All (T. Ohara); Tomakomai Children’s Clinic, Tomakomai (S. Fujibayashi); 11 genome segments of strain To14-0 (the representative Steel Memorial Muroran Hospital, Muroran, Japan (Y. Tahara); G8P[8] strain in this study) exhibited the highest nucleotide Japanese Red Cross Urakawa Hospital, Urakawa, Japan identity to human G8P[8] strains isolated in Southeast Asia (N. Kubo); Nakata Pediatric Clinic, Sapporo (S. Nakata); Sapporo in 2014 (represented by strain RVN1149 from Vietnam Hokushin Hospital, Sapporo (Y. Higashidate); National Institute of [>99.4% nt identity] and NP-130 from Thailand [>99.5% Infectious Diseases, Tokyo, Japan (Y. Fujii, K. Katayama) nt identity]) (6,7) (Table 2). This finding suggests that the strains share a common G8P[8] origin. DOI: https://dx.doi.org/10.3201/eid2306.160038

968 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Rotavirus G8P[8] Outbreak Strain, Japan, 2014

Figure 1. Distribution of rotavirus samples and their G/P genotypes in Hokkaido Prefecture (center map), Japan, 2014. The 4 locations from which the fecal samples were collected are shown on the map. Four hospitals (Tomakomai City Hospital, Japanese Red Cross Urakawa Hospital, Steel Memorial Muroran Hospital, and Sapporo Hokushin Hospital), and 2 clinics (Nakata Pediatric Clinic [Sapporo] and Tomakomai Children’s Clinic) participated in the study. Map at left shows location of Hokkaido in Japan (gray shading).

The VP7 gene of rotavirus strain To14-0 shared the In contrast, the VP1 and NSP4 genes of To14-0 highest nucleotide identity with the VP7 genes of human were only distantly related to those of the DS-1–like P[8] G8P[8] strains from Southeast Asia, including strains strains isolated in Asia (e.g., SKT-109, NT004, and LS- RVN1149 and NP-130 (99.4% and 99.7% identity, respec- 04), including the strains isolated in this study (e.g., To14- tively), and it shared slightly lower identity to the VP7 gene 41) (Table 2; online Technical Appendix Figure, panels of human strain 04-97s379 (97.8%) from Taiwan, which is A and I). The To14-0 VP1 gene shared high nucleotide speculated to be of bovine origin (8) (Figure 2). The VP7 identity with the VP1 genes of human G10P[14] strain genes of other G8 strains isolated in Japan were more dis- PR457 from Italy (98.1%), which are probably the result tantly related to the To14-0 VP7 gene (e.g., human AU109 of independent zoonotic transmissions (13). The To14- and bovine strains shared 89.5% and 81.9%–85.1%, re- 0 NSP4 gene shared high nucleotide identity with the spectively, with To14-0) (9). The VP7 genes of the human NSP4 genes of human strains BSGH38 from India G8 strains prevailing in Africa were also distantly related (96.7%) and the caprine G6P[1] strain GO34 from Ban- (<90% nt identity) to the VP7 gene of To14-0. gladesh (96.0%) (14). Among the 11 To14-0 genome segments, 8 (VP2–VP4, VP6, nonstructural protein [NSP] 1–3, and NSP5) were Conclusions highly similar to those of the human DS-1–like P[8] strains The clinical characteristics recorded for patients infected that have been isolated in Asia since 2012 (e.g., SKT-109, with G8P[8] rotaviruses and those infected with non- NT004, and LS-04) (10–12), including the strains isolated G8P[8] rotaviruses did not differ (Table 1). Our findings in this study (e.g., To14-41) (Table 2; online Technical Ap- suggest that the severity of gastroenteritis caused by newly pendix Figure, panels B–H, J). In addition, the VP6 and emerging G8P[8] rotaviruses could possibly be attenuated NSP5 genes of the strains isolated in this study were also by 1) the existence of VP7/VP4 genotype cross-reactive highly similar to those of human G2P[4] strains circulat- (heterotypic) protective responses; 2) protective immunity ing in South Korea (strain CAU15-11) and Thailand (strain associated with other segments, such as VP6 and NSP4

NP-M51) (12). (3,15); or 3) both of these factors combined.

Table 1. Characteristics of patients infected with (G8P[8] and non-G8P[8] rotavirus, Tomakomai, Hokkaido Prefecture, Japan, 2014 Patients infected with rotavirus Patient characteristic G8P[8], n = 42 Non-G8P[8], n = 42 p value* Sex, no. 1.0 M 24 24 F 18 18 Median age, y 1.8 2.0 0.50 Vaccinated, no (%) 2 (4.8) 1 (2.4) 0.56 Clinical symptoms, mean  SD Duration of fever, d 1.3 ± 1.2 1.7 ± 1.4 0.13 Duration of diarrhea, d 2.4 ± 1.6 2.5 ± 3.2 0.51 Diarrhea episode in 24 h, maximum no. 3.6 ± 3.2 4.3 ± 4.2 0.69 Duration of vomiting, d 1.4 ± 1.2 1.8 ± 2.9 0.79 Vomiting episodes in 24 h, maximum no. 3.6 ± 3.8 3.7 ± 3.5 0.98 Admitted to hospital, no. (%) 18 (42.9%) 16 (38.1%) 0.66 *Categorical data were analyzed by using 2 tests; continuous data were analyzed by using Student’s t-tests or Mann-Whitney U tests.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 969

DISPATCHES

Table 2. Genotype constellations and nucleotide identities of strains closely related to To14-0, the representative G8P[8] rotavirus strain used in a study of the clinical and molecular features of a G8P[8] rotavirus outbreak strain, Hokkaido Prefecture, Japan, 2014 Strain (genotype representative from Genotype constellations and nucleotide identities (%), by gene* study/country) VP7 VP4 VP6 VP1 VP2 VP3 NSP1 NSP2 NSP3 NSP4 NSP5 Human/To14-0 (G8P[8] in study) G8 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) Human/VNM/RVN1149/2014/G8P[8] G8 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 (G8P[8] in Vietnam) (99.4) (99.8) (99.8) (99.6) (99.6) (99.8) (100) (99.9) (99.5) (99.9) (99.5) Human/THA/NP-130/2014/G8P[8] G8 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 (G8P[8] in Thailand) (99.7) (99.9) (99.9) (99.9) (99.8) (99.9) (99.9) (99.7) (99.8) (99.7) (99.5) Human/THA/KKL-17/2013/G8P[8] G8 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 (G8P[8] in Thailand) (99.5) (99.8) (91.5) (99.9) (99.9) (99.9) (99.9) (99.8) (99.9) (99.7) (99.2) Human/THA/LS-04/2013/G2P[8] G2 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 (DS-1–like G2P[8] in Thailand) (–) (99.8) (99.5) (86.0) (99.8) (97.4) (99.9) (99.8) (98.6) (95.5) (99.7) Human/THA/SKT-109/2013/G1P[8] G1 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 (DS-1–like G1P[8] in Thailand) (–) (99.7) (98.7) (86.1) (99.9) (99.8) (99.8) (85.8) (99.8) (94.5) (99.2) Human/To14-41 (DS-1–like G1P[8] in G1 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 study) (–) (98.8) (98.3) (86.3) (99.4) (99.4) (99.5) (85.5) (99.6) (93.5) (98.8) Human/JPN/NT004/2012/G1P[8] G1 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 (DS-1–like G1P[8] in Japan) (–) (99.0) (98.4) (86.3) (99.7) (99.5) (99.5) (85.6) (99.6) (94.5) (99.2) Human/THA/NP-M51/2013/G2P[4] G2 P[4] I2 R2 C2 M2 A2 N2 T2 E2 H2 (G2P[4] in Thailand) (–) (–) (99.5) (86.0) (99.1) (97.5) (99.0) (85.4) (98.7) (95.5) (99.7) Human/KOR/CAU15–11/2015/G2P[4] G2 P[4] I2 R2 C2 M2 A2 N2 T2 E2 H2 (G2P[4] in South Korea) (–) (–) (99.5) (85.9) (99.2) (97.4) (94.0) (84.8) (98.1) (95.3) (98.6) Human/MU14114 (G2P[4] in study) G2 P[4] I2 R2 C2 M2 A2 N2 T2 E2 H2 (–) (–) (97.0) (85.3) (96.9) (86.4) (95.8) (85.5) (96.7) (95.6) (98.3) Human/ITA/PR457/2009/G10P[14] G10 P[14] I2 R2 C2 M2 A11 N2 T6 E2 H3 (–) (–) (93.0) (98.1) (82.3) (86.6) (–) (86.4) (–) (94.3) (–) Caprine/BGL/GO34/1999/G6P[1] G6 P[1] I2 R2 C2 M2 A11 N2 T6 E2 H3 (–) (–) (94.9) (86.3) (83.4) (86.8) (–) (86.6) (–) (96.0) (–) Human/COD/DRC88/2003/G8P[8] G8 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 (G8P[8] in Africa) (86.0) (98.3) (96.3) (85.6) (97.4) (97.0) (96.4) (85.8) (97.1) (90.1) (97.4) Human/MWI/QEC287/2006/G8P[8] G8 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 (G8P[8] in Africa) (85.5) (98.3) (96.3) (85.6) (97.8) (96.0) (96.5) (86.9) (96.6) (89.2) (97.6) Human/MWI/QEC289/2006/G8P[8] G8 P[8] I2 R2 C2 M2 A2 N2 T2 E2 H2 (G8P[8] in Africa) (85.5) (97.5) (96.3) (85.5) (97.8) (95.9) (96.5) (86.9) (96.6) (89.2) (97.6) *Purple indicates G8 genotype; green indicates Wa-like genome segments; red indicates DS-1–like genome segments; bold indicates nucleotide identities >99.0%; – indicates percentages not calculated because the genotype was different that for To14-0.

The VP7 genes of the human G8P[8] strains isolated The predominance of novel DS-1–like G8P[8] strains in this study and in Southeast Asia appear to have a close noted in this study indicates that these strains are suffi- relationship with bovine strains from Asia but not from Ja- ciently adapted to humans to sustain human-to-human pan, and the VP7 gene of human G8 or bovine G8 strains transmission in an industrialized country. This finding previously isolated in Japan are distantly related to them. suggests that these G8P[8] rotavirus strains could spread Therefore, the VP7 genes in the G8P[8] strains from this to other regions in the near future. Continuing surveil- study may have originated from a bovine strain from Asia. lance is required to monitor the circulating wild-type As with the VP7 genes, the VP1 and NSP4 genes are also strains, and rotavirus genotype constellations and clini- assumed to have been derived from artiodactyl strains. cal information must be analyzed to understand rotavirus Eight genome segments (VP2–VP4, VP6, NSP1– virulence in humans. NSP3, and NSP5) of the human G8P[8] strains isolated in This study was supported in part by a Japan Society for the this study and from Southeast Asia are closely related to Promotion of Science Grant-in-Aid for Scientific Research C those of the DS-1–like P[8] strains that have emerged and (grant no. 15K09693); a commissioned project for Research spread in Japan and other countries of Asia since 2012 (Ta- on Emerging and Re-emerging Infectious Diseases from the ble 2). Therefore, these 8 genome segments of the G8P[8] Japanese Ministry of Health, Labor and Welfare (to K.K.); strains from this study may be derived from the DS-1–like and Japan Agency for Medical Research and Development P[8] strains in Asia. (to K.K.). For the reasons we have stated, the G8P[8] strains isolated in this study were speculated to be formed outside of Dr. Kondo is a pediatrician in Department of Pediatrics, Japan by multiple reassortment events between the DS-1– Sapporo Medical University School of Medicine. His primary like P[8] strains and bovine strains in Asia. The resulting research interests are molecular biology and the epidemiology strain was probably recently introduced into Japan. of rotaviruses.

970 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Rotavirus G8P[8] Outbreak Strain, Japan, 2014

Figure 2. Phylogenetic analysis of the viral protein 7 gene of G8 rotavirus strains used in a study of the clinical and molecular features of a G8P[8] rotavirus outbreak strain, Hokkaido Prefecture, Japan, 2014. Closed circle indicates the G8P[8] rotavirus strain from Hokkaido; open circles indicate human G8P[8] strains from Southeast Asia; and closed boxes indicate other strains from Japan. A Tamura 3-parameter model was used for the maximum-likelihood method. Bootstrap values are shown at the branch nodes (values of <80% are not shown). Scale bar indicates nucleotide substitutions per site.

References 4. Tsugawa T, Tatsumi M, Tsutsumi H. Virulence-associated genome 1. Matthijnssens J, Ciarlet M, McDonald SM, Attoui H, Bányai K, mutations of murine rotavirus identified by alternating serial Brister JR, et al. Uniformity of rotavirus strain nomenclature passages in mice and cell cultures. J Virol. 2014;88:5543–58. proposed by the Rotavirus Classification Working Group (RCWG). http://dx.doi.org/10.1128/JVI.00041-14 Arch Virol. 2011;156:1397–413. http://dx.doi.org/10.1007/ 5. Masuda T, Nagai M, Yamasato H, Tsuchiaka S, Okazaki S, s00705-011-1006-z Katayama Y, et al. Identification of novel bovine group A rotavirus 2. Nakagomi T, Doan YH, Dove W, Ngwira B, Iturriza-Gómara M, G15P[14] strain from epizootic diarrhea of adult cows by de novo Nakagomi O, et al. G8 rotaviruses with conserved genotype sequencing using a next-generation sequencer. Vet Microbiol. constellations detected in Malawi over 10 years (1997–2007) display 2014;171:66–73. http://dx.doi.org/10.1016/j.vetmic.2014.03.009 frequent gene reassortment among strains co-circulating in humans. 6. Hoa-Tran TN, Nakagomi T, Vu HM, Do LP, Gauchan P, J Gen Virol. 2013;94:1273–95. http://dx.doi.org/10.1099/vir.0.050625-0 Agbemabiese CA, et al. Abrupt emergence and predominance in 3. Fujii Y, Shimoike T, Takagi H, Murakami K, Todaka-Takai R, Vietnam of rotavirus A strains possessing a bovine-like G8 Park Y, et al. Amplification of all 11 RNA segments of group A on a DS-1–like background. Arch Virol. 2016;161:479–82. rotaviruses based on reverse transcription polymerase chain reaction. http://dx.doi.org/10.1007/s00705-015-2682-x Microbiol Immunol. 2012;56:630–8. http://dx.doi.org/10.1111/ 7. Tacharoenmuang R, Komoto S, Guntapong R, Ide T, Sinchai P, j.1348-0421.2012.00479.x Upachai S, et al. Full genome characterization of novel DS-1–like

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 971 DISPATCHES

G8P[8] rotavirus strains that have emerged in Thailand: 12. Komoto S, Tacharoenmuang R, Guntapong R, Ide T, Tsuji T, reassortment of bovine and human rotavirus gene segments in Yoshikawa T, et al. Reassortment of human and animal rotavirus gene emerging DS-1–like intergenogroup reassortant strains. PLoS One. segments in emerging DS-1–like G1P[8] rotavirus strains. PLoS One. 2016;11:e0165826. http://dx.doi.org/10.1371/journal.pone.0165826 2016;11:e0148416. http://dx.doi.org/10.1371/journal.pone.0148416 8. Wu FT, Bányai K, Wu HS, Yang DC, Lin JS, Hsiung CA, et al. 13. Medici MC, Tummolo F, Bonica MB, Heylen E, Zeller M, Identification of a G8P[14] rotavirus isolate obtained from Calderaro A, et al. Genetic diversity in three bovine-like human a Taiwanese child: evidence for a relationship with bovine G8P[14] and G10P[14] rotaviruses suggests independent rotaviruses. Jpn J Infect Dis. 2012;65:455–7. http://dx.doi.org/ interspecies transmission events. J Gen Virol. 2015;96:1161–8. 10.7883/yoken.65.455 http://dx.doi.org/10.1099/vir.0.000055 9. Agbemabiese CA, Nakagomi T, Doan YH, Nakagomi O. 14. Ghosh S, Alam MM, Ahmed MU, Talukdar RI, Paul SK, Whole genomic constellation of the first human G8 rotavirus Kobayashi N. Complete genome constellation of a caprine group A strain detected in Japan. Infect Genet Evol. 2015;35:184–93. rotavirus strain reveals common evolution with ruminant and http://dx.doi.org/10.1016/j.meegid.2015.07.033 human rotavirus strains. J Gen Virol. 2010;91:2367–73. 10. Komoto S, Tacharoenmuang R, Guntapong R, Ide T, Haga K, http://dx.doi.org/10.1099/vir.0.022244-0 Katayama K, et al. Emergence and characterization of unusual 15. Burns JW, Siadat-Pajouh M, Krishnaney AA, Greenberg HB. DS-1–like G1P[8] rotavirus strains in children with diarrhea in Protective effect of rotavirus VP6-specific IgA monoclonal Thailand. PLoS One. 2015;10:e0141739. http://dx.doi.org/10.1371/ antibodies that lack neutralizing activity. Science. 1996;272:104–7. journal.pone.0141739 http://dx.doi.org/10.1126/science.272.5258.104 11. Fujii Y, Nakagomi T, Nishimura N, Noguchi A, Miura S, Ito H, et al. Spread and predominance in Japan of novel G1P[8] Address for correspondence: Takeshi Tsugawa, Department of Pediatrics, double-reassortant rotavirus strains possessing a DS-1–like | Sapporo Medical University School of Medicine, Minami 1-jo, genotype constellation typical of G2P[4] strains. Infect Genet Evol. 2014;28:426–33. http://dx.doi.org/10.1016/j. Nishi 16-chome, Chuo-ku, Sapporo, Hokkaido 060-8543, Japan; meegid.2014.08.001 email: [email protected] April 2015: Emerging Viruses • Reappearance of • Population Structure • Increased Risk for Chikungunya, Formerly and Antimicrobial Group B Streptococcus Called Dengue, in the Resistance of Invasive Sepsis in Young Americas Serotype IV Group B Infants Exposed to HIV, Streptococcus, Toronto, Soweto, South Africa, • Hantavirus Pulmonary Ontario, Canada 2004–2008 Syndrome, Southern Chile, 1995–2012 • Norovirus Genotype • Bat Coronavirus in Profiles Associated with Brazil Related to Foodborne Transmission, Appalachian Ridge 1999–2012 and Porcine Epidemic Diarrhea Viruses • Sequence Variability and Geographic • Tandem Repeat Distribution of Lassa Insertion in African Virus, Sierra Leone Swine Fever Virus, Russia, 2012 • Enterovirus D68 • Highly Pathogenic Infection, Chile, Avian Influenza A(H5N1) • Norovirus GII.21 in Spring 2014 Virus Infection among Children with • Animal-Associated Workers at Live Bird Diarrhea, Bhutan • Zika Virus Infection, Exposure to Rabies Markets, Bangladesh, Philippines, 2012 Virus among Travelers, 2009–2010 1997–2012 • Chikungunya • Deaths Associated with Outbreak, French • Evolution of Ebola Respiratory Syncytial Polynesia, 2014 Virus Disease from and Influenza Viruses Exotic Infection to • Avian Influenza among Persons A(H10N7) Virus– Global Health Priority, >5 Years of Age in Liberia, Mid-2014 Associated Mass HIV-Prevalent Area, Deaths among • Influenza A(H7N9) South Africa Harbor Seals Virus Transmission • Nairobi Sheep Disease between Finches • Hepatitis E Epidemic, Virus RNA in Ixodid and Poultry Biratnagar, Nepal, Ticks, China 2014 https://wwwnc.cdc.gov/eid/articles /issue/21/4/table-of-contents

972 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Seoul Virus Infection in Humans, France, 2014–2016

Jean-Marc Reynes, Damien Carli, increased levels of liver enzymes (Table 1). Hantavirus Jean-Baptiste Bour, Samir Boudjeltia, IgM and IgG were detected in an admission serum sample Anny Dewilde, Guillaume Gerbier, by using commercial ELISAs at a private clinical laborato- Timothée Nussbaumer, Véronique Jacomo, ry. Presence of IgM was confirmed at the French National Marie-Pierre Rapt, Pierre E. Rollin, Reference Center for Hantavirus (NRC; Lyon, France) by Alexandra Septfons using reference ELISA and indirect fluorescent antibody assays (5) (Table 2). We report detection of Seoul virus in 3 patients in France over Virus was detected by using molecular techniques a 2-year period. These patients accounted for 3 of the 4 Seoul as described (5). A partial SEOV small RNA sequence virus infections among 434 hantavirus infections (1.7%) re- (GenBank accession no. KX064269) was isolated from an ported during this time. More attention should be given to this virus in Europe where surveillance has been focused mostly admission sample by conducting a BLAST (https://blast. on Puumala and Dobrava-Belgrade hantaviruses. ncbi.nlm.nih.gov/Blast.cgi) search. Exposure to the virus was suspected to have occurred during building restoration work (Figure 1). eoul virus (SEOV), a hantavirus and the etiologic agent Case-patient 2 was a 22-year-old man from Erize- Sof a mild-to-moderate hemorrhagic fever with a renal Saint-Dizier, France (Figure 1), who was hospitalized 4 syndrome, is associated worldwide with brown rats (Rat- days after onset of symptoms during September 2014. His tus norvegicus), a commensal rodent that is found in all clinical and biologic characteristics were fever, thrombocy- human-inhabited locations (1). In Russia, South Korea, and topenia, and liver function disorders (Table 1). Hantavirus China, a wide range (few tens to few thousands) of human IgM and IgG were detected in an admission serum sample infections with SEOV are reported annually (with poten- by using commercial ELISAs at a public hospital clinical tial serologic cross-reactivity with co-circulating Hantaan laboratory. Results were confirmed at the NRC (Table 2). virus); otherwise, only a few countries have reported rare Virus was detected by using molecular techniques. A sporadic cases, mostly serologically confirmed (1–4). Four partial SEOV small RNA sequence was obtained from a human cases were recently described in the United King- serum sample, as shown by a BLAST search. dom and France (4–6). We report 3 patients in France, The pet rat (Rattus norvegicus) of case-patient 2, whose SEOV infections were confirmed by using molecu- bought ≈1 month before onset of symptoms, was consid- lar techniques, and identified through systematic surveil- ered the presumptive source of the virus. The animal was lance within a 24-month period. euthanized after consent of the patient was obtained. An identical partial SEOV small RNA sequence was obtained The Study from the liver of the animal. Case-patient 1 was a 27-year-old man from Dijon, France, Case-patient 3 was a 32-year-old man from Turckheim, who was hospitalized 3 days after onset of symptoms dur- France (Figure 1), who was hospitalized 4 days after onset of ing February 2014. His clinical and biologic characteristics symptoms during January 2016. His clinical characteristics were fever, mild renal syndrome, thrombocytopenia, and were severe fever, thrombocytopenia, liver disorders, myo- pericarditis, and renal syndrome that require hemodialysis Author affiliations: Institut Pasteur, Lyon, France (J.-M. Reynes, (Table 1). Hantavirus IgM and IgG were detected 6 days after D. Carli); Centre Hospitalier Universitaire Dijon-Bourgogne, Dijon, symptom onset by using commercial assays in a private clini- France (J.-B. Bour, S. Boudjeltia); Centre Hospitalier Universitaire cal laboratory. Results were confirmed at the NRC (Table 2). Lille, Lille, France (A. Dewilde); Direction Départementale de la Virus was detected by using molecular techniques re- Cohésion Sociale et de la Protection des Populations, Colmar, ported by Klempa et al. (7). A partial SEOV large RNA se- France (G. Gerbier); Hôpitaux Civils de Colmar, Colmar quence (GenBank accession no. KX064268) was obtained (T. Nussbaumer); Biomnis, Lyon (V. Jacomo); Centre Hospitalier, from an admission serum sample, as demonstrated by a Bar-le-Duc, France (M.-P. Rapt); Centers for Disease Control and BLAST search. Prevention, Atlanta, Georgia, USA (P.E. Rollin); Santé Publique This case-patient raised brown rats as a food source France, Saint-Maurice, France (A. Septfons) for his snakes and routinely captured and killed wild brown DOI: http://dx.doi.org/10.3201/eid2306.160927 rats that invaded his hen house. His breeding unit was the

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 973 DISPATCHES

Table 1. Characteristics of 3 patients infected with Seoul virus, France, 2014–2016* Characteristic† Patient 1 Patient 2 Patient 3 Signs and symptoms Fever Yes Yes Yes Weakness Yes No Yes Headache Yes Yes Yes Muscle or join pain No No Yes Chest pain No No Yes Abdominal pain No No Yes Nausea Yes No Yes Vomiting Yes No Yes Anorexia No No Yes Cough Yes No No Dyspnea No No Yes Anuria No No Yes Confusion Yes No Yes Laboratory results (reference range or value) Leukocytes, × 109 cells/L (4.0–10.0 × 109 cells/L) Reference Reference e D6: 19.4 Platelets, × 109/L (150–450 × 109/L) D4: 52 D4: 59 D4: 21 Hemoglobin, g/dL (13.0–17.0 g/dL) D6: 11.9 Reference e D13: 9.8 C-reactive protein, mg/L (<3.2 mg/L) D3: 82.0 D4: 26.0 D4: 155.0 LDH, IU/L (120–228 IU/L) D3: 999‡ NA D4: 876 AST, IU/L (15-37 IU/L) D3: 440 D5: 183 D4: 439 ALT, IU/L (21–72 IU/L) D9: 278 D5: 210 D4: 390 GGT, IU/L (15–85 IU/L) Reference D5: 87 D9: 722 Creatinine, mg/L (6–11 mg/L) D3: 16.4 Reference D8: 75.0 Microscopic hematuria NA No Yes Proteinuria, g/24 h (<0.3 g/24 h) NA NA D8: 3.82 *ALT, alanine aminotransferase; AST, aspartate aminotransferase; D, day after symptom onset; GGT -glutamyl transferase; LDH, lactate dehydrogenase; NA, not available; †Dysphagia, diarrhea, jaundice, sore throat, rash, bleeding manifestation, and loss of consciousness were not observed for these patients. ‡Patient 1 also had posttraumatic rhabdomyolysis: myoglobin level 869.5 g/mL at D3 (reference range 16.0–116.0 µg/mL) and creatine kinase level 9,444 g/mL at D5 (reference range 39–308 g/mL). likely source of human infection. Organs from 10 rats sam- infection with SEOV detected in 2014 (6). Using MEGA pled at this unit were positive for SEOV by nested reverse version 5.1 (8) and a generalized time-reversible model transcription PCR (7). A partial large RNA sequence ob- with a gamma distribution and 5 rate categories (accord- tained from a rat was identical to that obtained from the ing to the best fit substitution model proposed), we per- case-patient. Only 1 wild rat caught near the hen house was formed phylogenetic analysis of the small RNA coding sampled; results were negative for SEOV. sequence. This analysis confirmed that all virus strains We obtained complete small RNA coding do- detected were SEOV (Figure 2). main sequences (GenBank accession nos. KX064270– KX064275) from specimens from case-patients 2 and 3 Conclusions (samples were not available for case-patient 1); speci- Four hantaviruses (Puumala [PUUV], SEOV, Tula, and mens from the pet brown rat and raised brown rats sus- Nova viruses), have been detected in France; the first 3 pected to be sources of infection for case-patients 1 and viruses were associated with humans, and most human in- 2; and specimens from 2 wild brown rats suspected to fections were with PUUV (9). Human SEOV infection in sources of infection for a serologically confirmed human Europe was initially reported in 2013 (5); however, human

Table 2. Serologic results for hantaviruses for 3 patients with virologically confirmed infections with Seoul virus, France, 2014–2016* Diagnostic confirmation at NRC Additional test First-line diagnostic test, ELISA IFA at NRC‡ Patient Day of commercial ELISA (index value) IgM IgG Ig IgM no. sampling† IgM IgG SEOV PUUV SEOV PUUV SEOV PUUV PUUV 1 3 +§ (4.3) +§ (1.9) + – – – – – – 2 5 +¶ (7.9) +¶ (>200) + – E – + E – 29 ND ND + E + E + – ND 3 4 +§ (11.8) +§ (9.8) + + E – + – – 80 ND ND + – + + + + ND *E, equivocal; IFA, indirect fluorescent antibody; ND, not done; NRC, National Reference Center for Hantavirus (Lyon, France); PUUV, Puumala virus, SEOV, Seoul virus; +, positive; –, negative. †No. days after symptom onset. ‡Reascan Rapid Test (Reagena, Toivala, Finland). §Hantavirus IgG DxSelect and Hantavirus IgM DxSelect (Focus Diagnostics, Cypress, CA, USA). ¶Hantavirus Pool 1 Eurasia IgG and IgM (Euroimmun AG, Lübeck, Germany).

974 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Seoul Virus Infection in Humans, France, 2014–2016

Figure 1. Geographic distribution of Seoul virus (SEOV) infections among human and rats, France 2016. Gray shading area indicates area of France to which Puumala virus is endemic. Open green circles indicate SEOV serologically confirmed human infections with SEOV reported by Ragnaud et al. (10), Le Guenno (11), and Bour et al. (6); solid blue circles indicate virologically confirmed human infections with SEOV reported in this study (patients 1, 2, and 3); solid yellow square indicates virologically confirmed human infection with SEOV reported by Macé al. (5); solid red diamonds indicate virologically confirmed SEOV infections in brown rats reported in this study; and open blue diamonds indicate virologically confirmed SEOV infections in brown rats reported by Heyman et al. (12) and Dupinay et al. (13).

SEOV infections were previously suspected by serologic Most (83.4%) diagnostic tests in 2014 were requested analysis in France (6 cases were reported during 1977– for patients in the area of France to which PUUV is en- 1996). SEOV was also detected in rodents from several ar- demic. Consequently, some SEOV infections might have eas in France (Figure 1) (10–13). been be missed because hantavirus infection is rarely sus- Detection of SEOV in 3 case-patients in our study and pected outside this area. Samples with positive results for a recent report of a human infected with SEOV (6) within a SEOV are sent to NRC for diagnostic confirmation and 2-year period indicate that SEOV infections in humans are surveillance purposes. The 3 case-patients we report ini- not uncommon in France. These infections accounted for tially showed positive results by ELISA at local labora- 4 (1.7%) of 234 cases of hantavirus infection, mostly with tories but then showed negative results by a ReaScan PUUV, detected serologically or virologically during the Puumala IgM test (Reagena) at NRC (Table 2). Therefore, same period in France. Detection of such cases might be SEOV infections were probably not detected initially by caused by improvements in detection of SEOV infections, the 11 local laboratories who used the PUUV IgM Rapid rather than by emergence of SEOV infections (SEOV anti- Test, which led to underestimation of SEOV infections in gen in serologic assays and molecular detection have been humans in France. used at NRC since 2012). This negative result could have also occurred in other Some infections with SEOV are probably missed in countries in Europe that used the same test. Consequently, France. Commercials kits are used for hantavirus serologic use of a pan hantavirus serologic assay is preferred. Fur- diagnosis by 15 public hospital or private clinical labo- thermore, PUUV was detected by molecular techniques ratories in France. Eleven of these hospitals used a POC for most PUUV-infected patients during the acute phase Puumala IgM rapid test (Reagena, Toivala, Finland). Four of the disease (14). Thus, although there are no similar other hospitals used IgM and IgG ELISA kits with mix- data for SEOV, to avoid misdiagnosis, we suggest us- tures of recombinant antigens, including those from SEOV ing molecular diagnostic tests to more specifically detect or Hantaan virus strains. hantavirus infections.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 975 DISPATCHES

Figure 2. Phylogenetic tree based on the complete small RNA nucleotide coding sequences of Seoul virus (SEOV) strains isolated from 3 patients and rodents in contact with the patients infected with SEOV, France 2014–2016, and representative strains of SEOV and other hantavirus species. Triangles indicate sequences of strains detected in wild brown rats (this study) associated with a serologically confirmed human infection reported by Bour A et al. (6); squares indicate sequences of strains detected in case-patient 2 and in his pet brown rat; and circles indicate sequences of strains detected in case-patient 3 and 1 of his farmed brown rats. Bootstrap percentages >70% (from 500 resamplings) are indicated at each node; GenBank accession numbers are indicated for reference strains.Scale bar indicates nucleotide substitutions per site.

Furthermore, epidemiologic investigations showed that Emerg Infect Dis. 1997;3:205–11. http://dx.doi.org/10.3201/ several animal traders in Europe sold pet rats or food rats eid0302.970218 2. Kariwa H, Yoshimatsu K, Arikawa J. Hantavirus infection in to case-patients 2 and 3. These investigations also identified East Asia. Comp Immunol Microbiol Infect Dis. 2007;30:341–56. weak traceability of rat batches, absence of health-monitor- http://dx.doi.org/10.1016/j.cimid.2007.05.011 ing data for rats, and no information for the 3 case-patients 3. Knust B, Rollin PE. Twenty-year summary of surveillance for on zoonotic risks for infection. Awareness of rat owners, human hantavirus infections, United States. Emerg Infect Dis. 2013;19:1934–7. http://dx.doi.org/10.3201/eid1912.131217 traceability, and health monitoring of these animals, as per- 4. Jameson LJ, Taori SK, Atkinson B, Levick P, Featherstone CA, formed for laboratory rats (15), should be improved. van der Burgt G, et al. Pet rats as a source of hantavirus in England and Wales, 2013. Euro Surveill. 2013;18:20415. Acknowledgments 5. Macé G, Feyeux C, Mollard N, Chantegret C, Audia S, Rebibou JM, et al. Severe Seoul hantavirus infection in a pregnant We thank Christine Manson and David Reveille for sampling woman, France, October 2012. Euro Surveill. 2013;18:20464. brown rats. 6. Bour A, Reynes JM, Plaisancie X, Dufour JF. Seoul hantavirus infection-associated hemorrhagic fever with renal syndrome in The Centre National de Référence des Hantavirus was supported France: a case report [in French]. Rev Med Interne. 2016;37:493–6. by Santé Publique France. http://dx.doi.org/10.1016/j.revmed.2015.10.343 7. Klempa B, Fichet-Calvet E, Lecompte E, Auste B, Aniskin V, Dr. Reynes is a veterinarian and a medical virologist at the Meisel H, et al. Hantavirus in African wood mouse, Guinea. Institut Pasteur, Lyon, France. His research interests focus on Emerg Infect Dis. 2006;12:838–40. http://dx.doi.org/10.3201/ zoonotic infectious diseases. eid1205.051487 8. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. MEGA5: molecular evolutionary genetics analysis using maximum References likelihood, evolutionary distance, and maximum parsimony 1. Clement J, Heyman P, McKenna P, Colson P, Avsic-Zupanc T. methods. Mol Biol Evol. 2011;28:2731–9. http://dx.doi.org/ The hantaviruses of Europe: from the bedside to the bench. 10.1093/molbev/msr121

976 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Seoul Virus Infection in Humans, France, 2014–2016

9. Centre National de Reference Hantavirus, Institut Pasteur. 14. Lagerqvist N, Hagström Å, Lundahl M, Nilsson E, Juremalm M, Annual activity report, 2015 [in French] [cited 2017 Feb 27]. Larsson I, et al. Molecular diagnosis of Puumala virus-caused https://www.pasteur.fr/fr/sante-publique/CNR/les-cnr/hantavirus/ hemorrhagic fever with renal syndrome. J Clin Microbiol. rapports-d-activite 2016;54:1335–9. http://dx.doi.org/10.1128/JCM.00113-16 10. Ragnaud JM, Lamouliatte H, Paix MA, Fleury H, Roy J, 15. Mähler Convenor M, Berard M, Feinstein R, Gallagher A, Magnol R. Hemorrhagic fever with renal syndrome: a case with Illgen-Wilcke B, Pritchett-Corning K, et al.; FELASA jaundice [in French]. Gastroenterol Clin Biol. 1986;10:686. working group on revision of guidelines for health monitoring of 11. Le Guenno B. Hantaviruses [in French]. Médecine et Maladies rodents and rabbits. FELASA recommendations for Infectieuses. 1997;27:703–10. http://dx.doi.org/10.1016/ the health monitoring of mouse, rat, hamster, guinea pig and rabbit S0399-077X(97)80179-5 colonies in breeding and experimental units. 12. Heyman P, Plyusnina A, Berny P, Cochez C, Artois M, Zizi M, Lab Anim. 2014;48:178–92. http://dx.doi.org/10.1177/ et al. Seoul hantavirus in Europe: first demonstration of the virus 0023677213516312 genome in wild Rattus norvegicus captured in France. Eur J Clin Microbiol Infect Dis. 2004;23:711–7. http://dx.doi.org/10.1007/ Address for correspondence: Jean-Marc Reynes, Centre National de s10096-004-1196-3 Référence des Hantavirus, Unité de Biologie des Infections Virales 13. Dupinay T, Pounder KC, Ayral F, Laaberki MH, Marston DA, Emergentes, Institut Pasteur, Centre International de Recherche en Lacôte S, et al. Detection and genetic characterization of Seoul virus from commensal brown rats in France. Virol J. 2014;11:32. Infectiologie, 21 Ave Tony Garnier, 69 365 Lyon CEDEX 7, France; http://dx.doi.org/10.1186/1743-422X-11-32 email: [email protected]

April 2017: Emerging Viruses • Biologic Evidence Required for Zika Disease Enhancements by Dengue Antibodies Neurologic Complications of Influenza B Virus Infection in Adults, Romania • Implemen tation and Initial Analysis of a Laboratory-Based Weekly Biosurveillance System, Provence-Alpes-Côte d’Azur, France • Transmission of Hepatitis A Virus through Combined Liver–Small Intestine–Pancreas Transplantation • Influence of Referral Pathway on Ebola Virus Disease Case-Fatality Rate and Effect of Survival Selection Bias • Plasmodium malariae Prevalence and csp Gene Diversity, Kenya, • Reassortment of Influenza A Viruses in Wild Birds in Alaska 2014 and 2015 before H5 Clade 2.3.4.4 Outbreaks Incidence and Characteristics • Presence and Persistence of Zika Virus RNA in Semen, United of Scarlet Fever, South Korea, 2008–2015 Kingdom, 2016 • Markers of Disease Severity in Patients with Spanish Influenza in • Three Divergent Subpopulations of the Malaria Parasite the Japanese Armed Forces, 1919–1920 Plasmodium knowlesi • Molecular Identification of Spirometra erinaceieuropaei in Cases • Variation in Aedes aegypti Mosquito Competence for Zika Virus of Human Sparganosis, Hong Kong Transmission • Zika Virus Seroprevalence, French Polynesia, 2014–2015 * Persistent Arthralgia Associated with Chikungunya Virus Outbreak, US Virgin Islands, December 2014–February 2016 • Assessing Sensitivity and Specificity of Surveillance Case Definitions for Zika Virus Disease • Detection of Zika Virus in Desiccated Mosquitoes by Real-Time Reverse Transcription PCR and Plaque Assay • Surveillance and Testing for Middle East Respiratory Syndrome Coronavirus, Saudi Arabia, April 2015– February 2016 • Antiviral Drug-Resistant Influenza B • Outbreaks among Wild Birds and Domestic Poultry Caused Viruses Carrying H134N Substitution in Neuraminidase, Laos, February 2016 by Reassorted Influenza A(H5N8) Clade 2.3.4.4 Viruses, Germany, 2016 • Characterization of Highly Pathogenic Avian Influenza Virus A(H5N6), Japan, • Highly Pathogenic Avian Influenza A(H5N8) Virus in Wild November 2016 Migratory Birds, Qinghai Lake, China • Severe Thrombocytopenia After Zika • Design Strategies for Efficient Arbovirus Surveillance Virus Infection, Guadeloupe, 2016 • Typhus Group Rickettsiosis, Texas, 2003–2013 • Significant Decrease in Pertactin- • Detection and Molecular Characterization of Zoonotic Poxviruses Deficient Bordetella pertussis Isolates, Circulating in the Amazon Region of Colombia, 2014 Japan

https://wwwnc.cdc.gov/eid/articles/issue/23/4/table-of-contents

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 977 DISPATCHES

Central Nervous System Brucellosis Granuloma and White Matter Disease in Immunocompromised Patient

Mohammed Alqwaifly, Fahad S. Al-Ajlan, (5 mg 1×/d since 1993), levetiracetam (500 mg 2×/d for Hindi Al-Hindi, Abdulaziz Al Semari 5 mo), and phenytoin (200 mg every night for 1 mo). Neurologic examination showed left homonymous Brucellosis is a multisystem zoonotic disease. We report hemianopia, increased deep tendon reflexes in the left an unusual case of neurobrucellosis with seizures in an im- hemibody, and the Babinski sign on the left hallux. Initial munocompromised patient in Saudi Arabia who underwent laboratory test results, including those for complete blood renal transplantation. Magnetic resonance imaging of the brain showed diffuse white matter lesions. Serum and cere- count, erythrocyte sedimentation rate, C-reactive protein, brospinal fluid were positive for Brucella sp. Granuloma was and liver and renal profiles, were within references ranges. detected in a brain biopsy specimen. Results of serologic analysis for HIV and hepatitis B virus were negative. A standard agglutination tube (SAT) test result for uman brucellosis is a major zoonotic disease in Saudi Brucella spp. was positive (titer 1:320), and a 2-mercap- HArabia (1). This disease is caused by Brucella spp., toethanol test result for Brucella spp. agglutination was gram-negative bacteria usually transmitted through con- positive (titer 1:160). An ELISA showed antibodies against sumption of raw meat or unpasteurized dairy products (2). Brucella spp. in serum (titer 1:5,120). Cerebrospinal fluid Brucellosis is endemic to the Arabian Peninsula and coun- (CSF) had a leukocyte count of 21 (90% lymphocytes). tries bordering the Mediterranean Sea (3). Levels of protein, glucose, and lactate dehydrogenase in Neurobrucellosis occurs in 5%–10% of patients with CSF were within references ranges. brucellosis (4). The most frequent clinical manifestation Gram staining of a CSF sample and cultures for bacte- is meningoencephalitis (5). Mass lesions in the brain are ria, virus, fungi, and acid-fast bacilli (AFB) showed nega- uncommon (4). Intracerebral granuloma associated with tive results. Results of PCRs for AFB, cytomegalovirus, brucellosis had been reported in a community-acquired in- and JC polyomavirus were negative. fection (6). We report an unusual case of neurobrucellosis Serologic analysis of CSF showed Brucella IgG (titer and seizures in an immunocompromised patient. <1:20) and antibodies against Brucella (titer 1:320). Test results were negative for antibodies against Aspergillus, The Study Aspergillus galactomannan, Blastomyces, Borrelia, Coc- The patient was a 46-year-old Saudi woman who had cidia, Cryptococcus, Histoplasma, and Toxoplasma. chronic hepatitis C, end-stage renal disease of unde- An electroencephalogram showed sharp waves over termined etiology, and a renal transplant in 1993. She the right temporal region and continuous slow activity over reported a 5-month history of headaches and seizures. the right temporooccipital region. Magnetic resonance im- Seizures were usually preceded by epigastric pain and aging (MRI) of the brain showed diffuse T2/fluid-attenu- a sensation of nausea for few seconds, followed by left ated inversion recovery hyperintense white matter lesions arm posturing and loss of consciousness. She did not involving the right frontal, parietal and temporal lobes have fever, weight loss, or joint pain. She lived in a rural (Figure 1). No appreciable mass effect or enhancement area, was involved in animal husbandry, and consumed after administration of gadolinium was observed. Positive unpasteurized milk products. Her husband had been emission tomography of the brain showed hypometabolic treated for brucellosis. Her medications included myco- cerebral activity involving a large area of right cerebral phenolate mofetil (500 mg 2×/d since 1993), prednisone hemisphere. Magnetic resonance spectroscopy shows a low peak of n-acetyl aspartate (2.2 ppm). A brain biopsy specimen of cerebral cortex and su- Author affiliations: Qassim University College of Medicine, perficial white matter showed a moderate lymphoplasma- Qassim, Saudi Arabia (M. Alqwaifly); King Faisal Specialist cytic and focally histiocytic infiltrate that involved deep Hospital and Research Center, Riyadh, Saudi Arabia (F.S. Al-Ajlan, cortex, white matter, and leptomeninges. The histiocytic H. Al-Hindi, A. Al Semari) component formed small epithelioid granulomas that were DOI: http://dx.doi.org/10.3201/eid2306.161173 nonnecrotizing. The inflammatory reaction, including

978 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 CNS Brucellosis Granuloma and White Matter Disease

Figure 1. Magnetic resonance imaging of the brain of a 46-year-old immunocompromised woman with central nervous system brucellosis granuloma and white matter disease, Saudi Arabia. A) Axial T2 images showing hyperintensity in the right frontoparietal lobe and right temporal lobe. B) Axial fluid-attenuated inversion recovery (FLAIR) and C) coronal FLAIR images showing that hypersensitivity extends to U-fibers without involvement of the cortex. D) Gadolinium-enhanced image showing that no appreciable mass effect and no central or peripheral enhancement after administration of gadolinium were observed. Each image within each panel shows involvement in different levels of frontal, parietal, and temporal lobes. granulomas, was mainly perivascular with some angio- The patient received intravenous ceftriaxone (2 centric patterns and focal parenchymal involvement. The g every 12 h), oral doxycycline (100 mg every 12 h), white matter portion was heavily infiltrated by macro- oral rifampin (600 mg 1×/d), and trimethoprim/sulfa- phages. Reactive astrogliosis was prominent. There were methoxazole (1 tablet [160 mg/800 mg] every 12 h) for no morphologic signs of a specific etiology: no viral in- 2 wk. After discharge, she was receiving oral doxycy- clusions, and staining results microorganisms (AFB, fun- cline (100 mg every 12 h), rifampin (600 mg, 1×/d), tri- gi, Epstein-Barr virus, and JC polyomavirus) were nega- methoprim/sulfamethoxazole (1 tablet every 12 h), and tive (Figure 2). ciprofloxacin (500 mg every 12 h) for 6 mo: she was also A gram stain was initially negative for bacteria. At day receiving levitiracetam (750 mg 2×/d), carbamazepine 5, Brucella spp. were isolated from brain biopsy specimens. (200 mg 2×/d), mycophenolate mofetil (500 mg 2×/d), An antibiogram showed that the Brucella sp. was sensitive and prednisone (5 mg 1×/d). to gentamicin, streptomycin, tetracycline, trimethoprim/ Three months later, repeat MRI of the brain showed sulfamethoxazole, and rifampin. decreased T2 hyperintensity associated with volume loss

Figure 2. Histologic analysis of a brain biopsy specimen from a 46-year-old immunocompromised woman with central nervous system brucellosis granuloma and white matter disease, Saudi Arabia. A) Low magnification view of cerebral cortex showing infiltration by perivascular lymphocytes and histiocytes. Histiocytes form small nonnecrotizing granuloma (center) (original magnification ×100). B) High magnification view showing an angiocentric epithelioid granuloma cuffed by mature lymphocytes (original magnification ×200). Hemotoxylin and eosin stain.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 979 DISPATCHES and ex vacuo dilatation of the subjacent right lateral ven- features in a patient with seizures caused by chronic neuro- tricle. We did not observe any appreciable new lesions. brucellosis for 2.5 years (10). After 6 months of follow-up, her headaches had re- Granuloma is a pathogenesis of epilepsy (11). Soli- solved. However, she continued to have auras without ma- tary cysticercus granuloma and calcified lesion are 2 com- jor seizures. mon neuroimaging abnormalities in patients with epilepsy. Treatment for underlying cysticercosis does not cure epi- Conclusions lepsy (12). Seizures associated with central nervous system Neurobrucellosis can affect the central or peripheral ner- tuberculomas are often resolved after successful treatment vous systems and lead to diverse clinical syndromes (4). (13). The underlying pathogenesis for relapsing epilepsy in Diagnosis of neurobrucellosis depends on clinical manifes- neurocysticercosis is probably related to abnormal neurons tations, CSF findings suggestive of pinocytosis, high pro- and their arrangement within calcified nodules (13). The tein levels, low or standard glucose levels, and a positive epilepsy prognosis for neurobrucellosis is probably similar antibody titer for Brucella spp. Although the patient had to that for central nervous system neurocysticercosis (13). mild pleocytosis with a predominance of lymphocytes and A perivascular nonnecrotizing granuloma is a histo- high antibody titers against Brucella spp. in CSF, the CSF pathologic feature of neurobrucellosis. Neurocellosis gran- protein level was within the reference range. Antibodies uloma is a pathogenesis of refractory epilepsy. Our findings against Brucella spp. in CSF are usually an indication of indicate the need for suspecting neurobrucellosis as a cause neurobrucellosis. However, low levels of antibodies might of epilepsy and white matter disease in immunocompro- not be detected by SAT. In suspicious cases in which the mised patients in disease-endemic areas. SAT result is negative, SAT and a Coombs test, ELISA, Dr. Alqwaifly is an assistant professor of neurology at Qassim and PCR are helpful in making a diagnosis. University College of Medicine, Qassim, Saudi Arabia. 77The clinical−radiologic correlation for neurobrucel- His research interests are central nervous system infections, losis ranges from uneventful results for imaging studies, neuropathy, and movement disorders. despite positive clinical findings, to imaging abnormalities (3). Neurobrucellosis with a focal brain mass has been References rarely observed in imaging studies (7,8). 1. Kiel FW, Khan MY. Brucellosis in Saudi Arabia. Soc Sci Med. Radiologic results in this case suggested an infectious 1989;29:999–1001. http://dx.doi.org/10.1016/0277-9536(89) disease, autoimmune disease, or malignancy in an immu- 90056-7 2. Adams LG. The pathology of brucellosis reflects the outcome of nocompromised patient. Because we deemed it necessary the battle between the host genome and the Brucella genome. to exclude other conditions, such as progressive multifocal Vet Microbiol. 2002;90:553–61. http://dx.doi.org/10.1016/ leukoencephalopathy or lymphoma, we performed a brain S0378-1135(02)00235-3 biopsy. The diagnosis was established by detecting anti- 3. Al-Sous MW, Bohlega S, Al-Kawi MZ, Alwatban J, McLean DR. Neurobrucellosis: clinical and neuroimaging correlation. bodies against a Brucella sp. in serum and CSF and con- AJNR Am J Neuroradiol. 2004;25:395–401. firmed by isolation of aBrucella sp. from brain tissue. 4. Shakir RA. Neurobrucellosis. Postgrad Med J. 1986;62:1077–9. We found that the patient had epilepsy and extensive http://dx.doi.org/10.1136/pgmj.62.734.1077 white matter changes secondary to brucellosis. She contin- 5. Türel O, Sanli K, Hatipoğlu N, Aydoğmuş C, Hatipoğlu H, Siraneci R. Acute meningoencephalitis due to Brucella: case ued to have auras without major seizures. MRI of the brain report and review of neurobrucellosis in children. Turk J Pediatr. showed abnormal results (prominent white matter disease 2010;52:426–9. and focal encephalomalacia). Inflammation can cause per- 6. Sohn AH, Probert WS, Glaser CA, Gupta N, Bollen AW, Wong JD, manent cellular biochemical dysfunction, which can lead et al. Human neurobrucellosis with intracerebral granuloma caused by a marine mammal Brucella spp. Emerg Infect Dis. 2003;9:485– to electrically irritable tissue and parenchymal damage 8. http://dx.doi.org/10.3201/eid0904.020576 despite successful treatment. This finding might explain 7. Martínez-Chamorro E, Muñoz A, Esparza J, Muñoz MJ, the persistency of brain lesion. Appropriate antimicrobial Giangaspro E. Focal cerebral involvement by neurobrucellosis: therapy can eliminate the infection. pathological and MRI findings. Eur J Radiol. 2002;43:28–30. http://dx.doi.org/10.1016/S0720-048X(01)00390-4 Longitudinal studies of white matter hyperintensities 8. Erdem M, Namiduru M, Karaoglan I, Kecik VB, Aydin A, caused by vascular, noninfectious, infectious, and inflam- Tanriverdi M. Unusual presentation of neurobrucellosis: a solitary matory conditions showed white matter hyperintensities intracranial mass lesion mimicking a cerebral tumor: a case of over time despite effective treatment. Fincham et al. re- encephalitis caused by Brucella melitensis. J Infect Chemother. 2012;18:767–70. http://dx.doi.org/10.1007/s10156-011-0365-4 ported that white matter changes in neurobrucellosis were 9. Fincham RW, Sahs AL, Joynt RJ. Protean manifestation of nervous sequelae of demyelination, as confirmed by the pathologic system brucellosis. JAMA. 1963;184:269–76. http://dx.doi. analysis (9). We believe that unresolved white matter hy- org/10.1001/jama.1963.03700170061009 perintensities in this patient were a sequela of the inflam- 10. Yilmaz M, Ozaras R, Ozturk R, Mert A, Tabak F, Aktuglu Y. matory process. A case report documented similar clinical Epileptic seizure: an atypical presentation in an adolescent boy

980 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 CNS Brucellosis Granuloma and White Matter Disease

with neurobrucellosis. Scand J Infect Dis. 2002;34:623–5. 13. AlSemari A, Baz S, Alrabiah F, Al-Khairallah T, Qadi N, http://dx.doi.org/10.1080/00365540210147561 Kareem A, et al. Natural course of epilepsy concomitant 11. Sotelo-Morales J, García-Cuevas E, Rubio-Donnadieu F. Granuloma with CNS tuberculomas. Epilepsy Res. 2012;99:107–11. in cerebral parenchyma. A human model for the study of epilepsy http://dx.doi.org/10.1016/j.eplepsyres.2011.10.032 [in Spanish]. Gac Med Mex. 1989;125:31–5, discussion 36. 12. Sharma LN, Garg RK, Verma R, Singh MK, Malhotra HS. Seizure Address for correspondence: Mohammed Alqwaifly, Qassim recurrence in patients with solitary cystic granuloma or single University College of Medicine, Qassim 51391, Saudi Arabia: parenchymal cerebral calcification: a comparative evaluation. Seizure. 2013;22:840–5. http://dx.doi.org/10.1016/j.seizure.2013.07.001 email: [email protected]

July 2016: Zoonoses • Two Linked Enteroinvasive • High Incidence of Escherichia coli Chikungunya Virus and Outbreaks, Nottingham, Frequency of Viremic United Kingdom, Blood Donations during June 2014 Epidemic, Puerto Rico, USA, 2014 • Porcine Bocavirus Infection Associated with • Outbreak of Vibrio Encephalomyelitis in a parahaemolyticus Pig, Germany Sequence Type 120, • African Swine Fever Peru, 2009 Epidemic, Poland, • Clinical Manifestations 2014–2015 of Senecavirus A • Hepatitis E Virus in Infection in Neonatal Dromedaries, North and Pigs, Brazil, 2015 • Surveillance for Highly • Turtle-Associated East Africa, United Arab Pathogenic Avian Salmonellosis, United Emirates and Pakistan, Influenza Virus in Wild States, 2006–2014 1983–2015 Birds during Outbreaks • Pregnancy, Labor, and in Domestic Poultry, • Heatwave-Associated Minnesota, 2015 Delivery after Ebola Virus Vibriosis, Sweden and Disease and Implications Finland, 2014 • Highly Pathogenic Avian for Infection Control Influenza Viruses and in Obstetric Services, • Vesicular Disease Generation of Novel United States, 2015 in 9-Week-Old Pigs Reassortants, United Experimentally Infected States, 2014–2015 • Response to Middle East with Senecavirus A Respiratory Syndrome • Naturally Circulating Coronavirus, Abu Dhabi, Hepatitis A Virus in Olive United Arab Emirates, Baboons, Uganda 2013–2014 • Detection and Genomic • Current Guidelines, Characterization of Common Clinical Pitfalls, Senecavirus A, and Future Directions for • Infection with Possible Ohio, USA, 2015 Laboratory Diagnosis of Novel Parapoxvirus in Lyme Disease, United Horse, Finland, 2013 • Red Fox as a Sentinel for Blastomyces dermatitidis, • Tropheryma whipplei • Travel-Associated Rabies Ontario, Canada as a Cause of Epidemic in Pets and Residual Fever, Senegal, Rabies Risk, Western • Senecavirus A in Pigs, 2010–2012 Europe United States, 2015

https://wwwnc.cdc.gov/eid/articles/issue/22/7/table-of-contents

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 981 DISPATCHES

Severe Neurologic Disorders in 2 Fetuses with Zika Virus Infection, Colombia

Jorge Acosta-Reyes, Edgar Navarro, callosum, and atrophy of the cerebellum and brainstem in Maria José Herrera, Eloina Goenaga, fetuses with congenital Zika virus infection (6). This re- Martha L. Ospina, Edgar Parra, Marcela Mercado, port describes 2 fetuses examined after pregnancy termi- Pablo Chaparro, Mauricio Beltran, nation who had severe CNS defects attributed to maternal Maria Luz Gunturiz, Lissethe Pardo, Zika virus infection. Catalina Valencia, Sandra Huertas, Jorge Rodríguez, Germán Ruiz, Diana Valencia, The Cases Lisa B. Haddad, Sarah C. Tinker, Case 1 involved a 24-year-old pregnant resident of the Cynthia A. Moore, Hernando Baquero city of Barranquilla who had symptoms compatible with Zika virus disease, including fever and generalized We report the results of pathologic examinations of 2 fe- rash for 2 days, followed by edema and joint pain for 2 tuses from women in Colombia with Zika virus infection dur- weeks, with onset occurring at 5–6 weeks of gestation. ing pregnancy that revealed severe central nervous system Her medical history was notable because of a prior preg- defects and potential associated abnormalities of the eye, spleen, and placenta. Amniotic fluid and tissues from mul- nancy complicated by a stillbirth with Potter syndrome, tiple fetal organs tested positive for Zika virus. delivered 2 months before conception of the current pregnancy. Obstetric ultrasounds at 8 and 12 weeks of gestation showed a singleton fetus with no abnormali- n October 2015, Colombia confirmed its first case of ties; an ultrasound at 15 weeks of gestation demonstrat- Iautochthonous Zika virus infection (1). As of Novem- ed absence of the cranial vault (exencephaly) with brain ber 2016, ≈105,000 cases of symptomatic Zika virus dis- tissue loss, including absence of cerebral hemispheres ease have been reported, including 19,499 cases among (anencephaly sequence). Amniotic fluid sampled at 17 pregnant women. Although Zika virus infections typically weeks was positive for Zika virus by real-time reverse lead to comparatively benign symptoms relative to those transcription PCR (rRT-PCR). The fetus was at 20 of other arboviruses (2), in May 2016, the US Centers weeks of gestation when evaluated. for Disease Control and Prevention concluded that con- Case 2 involved a 15-year-old pregnant resident of genital Zika virus infection was the cause of the severe Bogotá who reported symptoms compatible with Zika central nervous system (CNS) defects observed in fetuses virus disease, including generalized rash for 2 days and and newborns of women infected with Zika virus during myalgia, after travel to a municipality with prevalent Zika pregnancy (3). Although the full spectrum of fetal effects virus infections, with onset occurring at 16–20 weeks of of congenital Zika virus infection is not known, Zika vi- gestation. She had an uncomplicated obstetric history with rus has been shown to cross the placental barrier, grow 1 prior live birth. Prenatal care started in week 26; an ul- in brain tissue of fetuses, infect progenitor neural cells, trasound at that time showed a fetus with abnormal clefts and increase neural cell death or attenuate their growth in the cerebral hemispheres of the brain (schizencephaly), (4,5). Observed CNS abnormalities include microcephaly, with cleft walls separated and filled with cerebrospinal flu- ventriculomegaly, cerebral calcifications, absent corpus id (open lip). Follow-up ultrasound indicated large regions Author affiliations: Universidad del Norte, Barranquilla, of parietal and temporal brain parenchymal loss. The fetus Colombia (J. Acosta-Reyes, E. Navarro, M.J. Herrera, was evaluated and Zika virus diagnosis confirmed at 27 H. Baquero); Secretaría de Salud de Barranquilla, Barranquilla weeks of gestation. (E. Goenaga); Instituto Nacional de Salud, Bogotá, Colombia Autopsies of fetal organs and placenta with anatomic (M.L. Ospina, E. Parra, M. Mercado, P. Chaparro, M. Beltrán, pathology and microscopic evaluation were performed for M.L. Gunturiz, L. Pardo); Clínica Colsanitas, Clínica Reina Sofía, both fetuses. Tissue samples were fixed in 10% neutral- Bogotá (C. Valencia, S. Huertas, J. Rodríguez, G. Ruiz); Centers buffered formalin and fixed in paraffin; 4-µm-thick cuts for Disease Control and Prevention, Atlanta, Georgia, USA were stained with hematoxylin and eosin for morphologic (D. Valencia, S.C. Tinker, C.A. Moore); Emory University School of analysis with light microscopy. RNA was extracted from Medicine, Atlanta (L.B. Haddad). fresh tissue by using a TRIzol plus RNA purification kit (Thermo Fisher Scientific, Waltham, MA, USA). We used DOI: http://dx.doi.org/10.3201/eid2306.161702

982 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Neurologic Disorders and Zika Infection, Colombia

Figure 1. Pathology findings for case 1, involving a fetus examined after pregnancy termination who had severe neurologic defects attributed to maternal Zika virus infection, Colombia. A) Remnant tissue of cerebral cortex showing a reduced neuroblast layer (dotted lines) and hemorrhagic foci (arrow). Hematoxylin and eosin (H&E) staining; original magnification ×40. B) Glial leptomeningeal heterotopy showing congestive blood vessels (arrowhead) and foci of glial heterotopia (arrows). H&E staining; original magnification ×10. C) Glial leptomeningeal heterotopy showing congestive blood vessels (arrowhead) and foci of glial heterotopia (arrows). H&E staining; original magnification ×40. rRT-PCR for the detection of Zika virus (NS5) and 1-step cerebrum and brain stem, with no microcalcifications, in- RT-PCR for the detection of the envelope protein–coding clusion bodies, or inflammatory infiltrate (Figure 1). For region (360 bp) were performed as previously described case 2, we observed microcalcifications in areas of neu- (7,8). Immunohistochemical markers were used in tissue rologic degeneration; a spinal cord slice showed neuropil from both fetuses. Molecular testing for syphilis, toxoplas- disruption with multiple calcifications, reactive gliosis, mosis, rubella, and cytomegalovirus was performed fol- and loss of the ascending and descending spinal cord lowing techniques described previously (9,10); 500-band tracts (Figure 2). resolution karyotype was also performed for both fetuses, These 2 cases of fetal Zika virus infection involved and ELISA testing for HIV was performed on maternal se- severe brain tissue damage and expand the spectrum of rum samples. CNS defects that might result from congenital Zika virus We documented notable abnormalities in the CNS, infection (11). Multiple tissues tested positive for Zika eye, spleen, and placenta for 1 or both fetuses (Table, virus in both cases, providing strong evidence of vertical https://wwwnc.cdc.gov/EID/article/23/1/16-1702-T1.htm). transmission. For case 2, the observed pattern of arrest in All tissues tested and amniotic fluid were positive for Zika neuronal migration and decrease in the pattern of matura- virus except eye and placenta from the fetus in case 2. Im- tion of the CNS is consistent with the potential effects of munohistochemical markers were negative. For both cases, Zika virus infection noted in previous reports (5). For case all tissue samples were negative for other congenital infec- 1, although Zika virus was demonstrated in multiple tis- tions (i.e., syphilis, toxoplasmosis, rubella, and cytomega- sues of the fetus, the birth defect identified (anencephaly) lovirus), high-resolution karyotype was normal, and mater- is not on the recognized spectrum of congenital infection. nal serum was HIV-negative. The timing of exposure and primary affected tissue type For case 1, microscopic examination demonstrated potentially implicate Zika virus in disruption of neural tube evidence of recent bleeding in the residual tissues of the closure in this case. However, the pathology evaluation

Figure 2. Pathology findings for case 2, involving a fetus examined after pregnancy termination who had severe neurologic defects attributed to maternal Zika virus infection, Colombia. A) Spinal cord slice showing neuropil disruption with multiple calcifications (arrowheads). B) Nerve showing disruptive changes of axons (Wallerian degeneration) (black star). C) Dorsal root ganglion showing spinal ganglion with satellitosis (arrow) and neuronophagia of ganglion cells (arrowhead). Hematoxylin and eosin staining; original magnification ×100.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 983 DISPATCHES was consistent with expected findings in any fetus with References anencephaly; therefore, congenital Zika virus might have 1. Pacheco O, Beltrán M, Nelson CA, Valencia D, Tolosa N, Farr SL, et al. Zika virus disease in Colombia—preliminary report. been coincidental in a fetus already predestined by other N Engl J Med. 2016;NEJMoa1604037. http://dx.doi.org/10.1056/ factors to have anencephaly. NEJMoa1604037 The identification of virus in tissues outside of the 2. Petersen LR, Jamieson DJ, Honein MA. Zika virus. N Engl J Med. brain suggests the need to consider other organs as targets 2016;375:294–5. https://dx.doi.org/10.1056/NEJMc1606769 3. Rasmussen SA, Jamieson DJ, Honein MA, Petersen LR. Zika virus for the virus. To improve knowledge of the spectrum of and birth defects—reviewing the evidence for causality. N Engl J outcomes among pregnancies affected by congenital Zika Med. 2016;374:1981–7. http://dx.doi.org/10.1056/NEJMsr1604338 virus and to identify interventions that might reduce ill- 4. Mlakar J, Korva M, Tul N, Popović M, Poljšak-Prijatelj M, ness among affected newborns, such as hearing and visual Mraz J, et al. Zika virus associated with microcephaly. N Engl J Med. 2016;374:951–8. http://dx.doi.org/10.1056/NEJMoa1600651 screening, it will be necessary to look for damage in organs 5. Tang H, Hammack C, Ogden SC, Wen Z, Qian X, Li Y, other than the brain among newborns of mothers with Zika et al. Zika virus infects human cortical neural progenitors and virus infection during pregnancy. The mothers of the fe- attenuates their growth. Cell Stem Cell. 2016;18:587–90. tuses in these 2 cases reported Zika virus disease symptoms http://dx.doi.org/10.1016/j.stem.2016.02.016 6. Moore CA, Staples JE, Dobyns WB, Pessoa A, Ventura CV, at different times during the pregnancy (early first trimes- Fonseca EB, et al. Characterizing the pattern of anomalies in ter and early second trimester), which highlights the po- congenital Zika syndrome for pediatric clinicians. JAMA Pediatr. tentially long window of vulnerability to adverse outcomes 2016; Nov 3 [Epub ahead of print]. http://dx.doi.org/10.1001/ among pregnant women with Zika virus infection. jamapediatrics.2016.3982 7. Faye O, Faye O, Diallo D, Diallo M, Weidmann M, Sall AA. Persistent viremia in maternal serum has been previ- Quantitative real-time PCR detection of Zika virus and ously reported (12–14); we observed positive results for evaluation with field-caught mosquitoes. Virol J. 2013;10:311. Zika virus molecular tests performed on fetal tissues and http://dx.doi.org/10.1186/1743-422X-10-311 amniotic fluid 15 week (case 1) and 10 weeks (case 2) af- 8. Faye O, Faye O, Dupressoir A, Weidmann M, Ndiaye M, Alpha Sall A. One-step RT-PCR for detection of Zika virus. J Clin ter symptoms of acute infection in the pregnant women. Virol. 2008;43:96–101. http://dx.doi.org/10.1016/j.jcv.2008.05.005 Although we did not test for replication-competent virus, 9. Bosma TJ, Corbett KM, O’Shea S, Banatvala JE, Best JM. PCR for the virus probably persists in different fetal tissues after detection of rubella virus RNA in clinical samples. J Clin primary infection because of replication in those tissues. Microbiol. 1995;33:1075–9. 10. Cortés LJ, Duque S, López MC, Moncada D, Molina D, The virus might also replicate in placental tissue (12,15), Gómez-Marín JE, et al. [Gene polymorphisms in the dihydrofolate although it is probably not the only location for viral repli- reductase (dhfr) and dihydropteroate synthase (dhps) genes and cation in the 2 cases described in this report, given that Zika structural modelling of the dhps gene in Colombian isolates of virus was identified in the placenta for only 1 of the cases. Toxoplasma gondii [in Spanish]. Biomedica. 2014;34:556–66. http://dx.doi.org/10.7705/biomedica.v34i4.2132 Given the epidemic levels of Zika virus disease being 11. Martines RB, Bhatnagar J, de Oliveira Ramos AM, Davi HP, reported in Colombia and the severe outcomes associated Iglezias SD, Kanamura CT, et al. Pathology of congenital Zika with congenital infection, we call attention to the impor- syndrome in Brazil: a case series. Lancet. 2016;388:898–904. tance of healthcare providers reporting cases of Zika virus http://dx.doi.org/10.1016/S0140-6736(16)30883-2 12. Driggers RW, Ho CY, Korhonen EM, Kuivanen S, Jääskeläinen AJ, disease for adequate public health surveillance. Recom- Smura T, et al. Zika virus infection with prolonged maternal mendations might include enhanced surveillance of acute viremia and fetal brain abnormalities. N Engl J Med. cases among pregnant women and stillbirths and abor- 2016;374:2142–51. http://dx.doi.org/10.1056/NEJMoa1601824 tions related to Zika virus disease during pregnancy, as 13. Bhatnagar J, Rabeneck DB, Martines RB, Reagan-Steiner S, Ermias Y, Estetter LB, et al. Zika virus RNA replication and well as adverse outcomes among live births. Collecting persistence in brain and placental tissue. Emerg Infect Dis. information on congenital Zika virus infection is impor- 2017;23:405–14. http://dx.doi.org/10.3201/eid2303.161499 tant to helping public health authorities effectively -ad 14. Jurado KA, Simoni MK, Tang Z, Uraki R, Hwang J, dress this new health threat and make evidence-based de- Householder S, et al. Zika virus productively infects primary human placenta-specific macrophages. JCI Insight. 2016; cisions, which will benefit Colombia and other countries 1:pii:e88461. https://dx.doi.org/10.1172/jci.insight.88461 throughout the world. 15. Meaney-Delman D, Oduyebo T, Polen KN, White JL, Bingham AM, Slavinski SA, et al. Prolonged detection of Zika Dr. Acosta-Reyes is a clinical epidemiologist and currently is virus RNA in pregnant women. Obstet Gynecol. 2016;128:724–30. the Director of the Department of Public Health in the School http://dx.doi.org/10.1097/AOG.0000000000001625 of Medicine at Universidad del Norte, Barranquilla, Colombia. His research interests include the epidemiology, immunology, Address for correspondence: Jorge Acosta-Reyes, Universidad del Norte, diagnosis, and prevention of arboviruses and maternal and Km 5 Vía Puerto Colombia, Barranquilla, Colombia; email: child health. [email protected]

984 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Domestic Pig Unlikely Reservoir for MERS-CoV

Emmie de Wit, Friederike Feldmann, Eva Horne, without the constraints associated with nonhuman primate Cynthia Martellaro, Elaine Haddock, studies, we explored the possibility of using the domestic Trenton Bushmaker, Kyle Rosenke, pig as an animal model of MERS-CoV infection. Atsushi Okumura, Rebecca Rosenke, Greg Saturday, Dana Scott, Heinz Feldmann The Study Comparison of the DPP4 nucleotide sequences of humans, We tested the suitability of the domestic pig as a model for dromedary camels, and domestic pigs showed that the por- Middle East respiratory syndrome coronavirus (MERS-CoV) cine DPP4 is identical to the dromedary camel DPP4 at the infection. Inoculation did not cause disease, but a low level 14 aa positions that have been shown to determine species of virus replication, shedding, and seroconversion were observed. Pigs do not recapitulate human MERS-CoV and are tropism (Table) (8,9). We investigated whether DPP4 is ex- unlikely to constitute a reservoir in nature. pressed in the pig respiratory tract by performing immunohistochemical staining on the nasal mucosa and lung tissue obtained from healthy pigs using an antibody against DPP4 s of March 10, 2017, a total of 1,917 cases of Middle (mouse monoclonal anti-DPP4 [CD26], clone OTI11D7, AEast respiratory syndrome coronavirus (MERS-CoV) 1:2,500; Origene Technologies, Inc., Rockville, MD, USA). infection and 684 fatalities have occurred (1). Despite the DPP4 expression was not observed in the nasal mucosa relatively large number of cases, little is known about the of healthy domestic pigs (Figure 1, panel A); in the lungs, disease pathology of MERS in humans (2). Our current un- abundant DPP4 expression was observed in type I and type derstanding of the pathogenesis of MERS-CoV is therefore II pneumocytes and submucosal glands (Figure 1, panel B), mostly based on data derived from studies in animal mod- suggesting MERS-CoV infection would be supported. els. Although the first animal model used to study MERS- We inoculated 2 groups of four 4–5-week-old farm CoV pathogenesis and test potential countermeasures be- pigs (Yorkshire cross; S&S Farms, Ramona, CA, USA) in- came available shortly after the discovery of MERS-CoV tranasally (1 mL/nostril) and intratracheally (5 mL) with a (3), all the animal models that have been developed so far 6 total dose of 10 tissue culture infectious dose 50 (TCID50) have drawbacks (4). Because of the host restriction con- of the hCoV-EMC/2012 isolate of MERS-CoV. A group ferred by the binding of the MERS-CoV spike protein to of 3 control pigs was mock inoculated with Dulbecco’s its receptor, dipeptidyl peptidase 4 (DPP4), small animal modified Eagle medium (DMEM); these pigs were housed models that are routinely used to conduct infectious disease in a separate room from the MERS-CoV–inoculated pigs research are not naturally susceptible to MERS-CoV in- to prevent cross-contamination. Animal experiments were fection. Although human DPP4-transgenic mouse models approved by the Institutional Animal Care and Use Com- have been developed, these do not completely recapitulate mittee of the Rocky Mountain Laboratories and conducted the disease pathology observed in humans. Nonhuman pri- by certified staff in an Association for Assessment and Ac- mate models recapitulate mild and moderate human disease creditation of Laboratory Animal Care International–ac- pathology; however, practical and ethical constraints limit credited facility according to the institution’s guidelines work with these models. for animal use; staff followed the guidelines and basic prin- The domestic pig (Sus domesticus) is used in infectious ciples in the US Public Health Service Policy on Humane disease research because of similarities between human Care and Use of Laboratory Animals and the Guide for the and pig anatomy, genetics, and physiology (5). MERS- Care and Use of Laboratory Animals. CoV was previously shown to replicate in porcine kidney After inoculation with MERS-CoV, none of the pigs cells, albeit less efficiently than in human kidney cells (6). showed clinical signs of disease, such as increased body In an effort to develop a MERS-CoV animal model that re- temperature or increased respiration, and bodyweight gain capitulates human disease better than small animal models was similar between MERS-CoV–inoculated and mock- inoculated pigs (Figure 2, panel A). We collected nose and Author affiliations: National Institutes of Health, Hamilton, throat swabs during clinical exams and analyzed them for Montana, USA (E. de Wit, F. Feldmann, E. Horne, C. Martellaro, the presence of viral RNA by quantitative reverse tran- E. Haddock, T. Bushmaker, K. Rosenke, R. Rosenke, G. Saturday, scription PCR (qRT-PCR) as described (10). Shedding of D. Scott, H. Feldmann); Columbia University, New York, New York, viral RNA from the nose and the throat increased from 1 USA (A. Okumura) day postinoculation (dpi) to 3 dpi in all MERS-CoV–in- DOI: http://dx.doi.org/10.3201/eid2306.170096 oculated animals, a sign that active replication occurred;

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 985 DISPATCHES

Table. Comparison of the amino acid residues shown to be essential in binding of Middle East respiratory syndrome coronavirus spike protein to DPP4 of human, dromedary camel, and domestic pig* DPP4, aa position Species 229 267 286 288 291 294 295 298 317 322 336 341 344 346 Human† N K Q T A L I H R Y R V Q I Dromedary camel‡ – – – V – – – – – – – – – – Domestic pig§ – – – V – – – – – – – – – – Mouse¶ – – – P – A R – – – T S – V *DPP4, dipeptidyl peptidase 4; –, no change from human DPP4. †GenBank accession no. AB451339. ‡GenBank accession no. KF574263. §GenBank accession no. NM214257. ¶Taken from previous publication (7). shedding was higher in the nose than in the throat (Figure On 4 and 11 dpi, we euthanized 4 MERS-CoV–in- 2, panel B). After 3 dpi, shedding of viral RNA decreased; oculated pigs and collected their tissues for virologic all nose swabs were negative by 11 dpi and all throat swabs and histologic analysis. Viral RNA could be detected by by 7 dpi. qRT-PCR in >1 respiratory tract tissue samples of all 4 We attempted virus propagation by inoculating VeroE6 MERS-CoV–inoculated pigs. However, viral loads were cells with the media used to resuspend nasal and throat swab low; infectious MERS-CoV could not be isolated from particulates and checking for the development of MERS- any tissues positive by qRT-PCR, and the distribution of CoV cytopathic effect. Infectious MERS-CoV was not re- viral RNA among tissues was inconsistent from pig to pig covered at any time postinoculation from any swab sample. (Figure 2 panel C). By 11 dpi, viral RNA could only be detected in the bronchial lymph node of 1 MERS-CoV– inoculated pig (Figure 2 panel C); all the tissues examined from other MERS-CoV–inoculated pigs were negative by this time. Viral RNA could not be detected in any of the extrarespiratory tissues tested, such as heart, liver, spleen, kidney, adrenal gland, duodenum, ileum, transverse colon, or urinary bladder, on 4 dpi or 11 dpi (data not shown). Histologic analysis did not reveal any lesions consistent with MERS-CoV infection in any of the collected tissues, including those of the respiratory tract. We performed immunohistochemical staining with an antibody specific for MERS-CoV on tonsil, trachea, bronchial lymph node, and right and left lower lung lobe of all pigs, as well as other tissues that tested positive for viral RNA by qRT-PCR. MERS-CoV antigen could not be detected in any of these tissues. Serum samples collected on the day of euthanasia were tested for the presence of antibodies against MERS-CoV spike protein 1 (S1) by ELISA. By 11 dpi, antibodies directed against MERS-CoV S1 could be detected in all 4 pigs (Figure 2, panel D).

Conclusions Recently, Vergara-Alert et al. showed MERS-CoV shed- 7 ding in pigs inoculated with 10 TCID50 of MERS-CoV and suggested that pigs could play a role as a reservoir for the circulation of MERS-CoV (12). In our hands, Figure 1. Dipeptidyl peptidase (DPP) 4 expression in the domestic pigs inoculated with a 10-fold lower infectious dose of pig respiratory tract. Tissues were stained by using a cross- MERS-CoV were also successfully infected, but the low reactive mouse monoclonal antibody against DPP4 (CD26, clone amount of virus replication in and shedding from the re- OTI11D7, 1:2,500; Origene Technologies, Inc., Rockville, MD, USA). DPP4 expression was absent in the nasal mucosa (A) spiratory tract implies that the pig is unlikely to play a but present in lung tissue (B) of healthy domestic pigs. Original profound role as an intermediate host for MERS-CoV magnification: nasal mucosa ×40; lung ×200. in nature.

986 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Domestic Pig Unlikely Reservoir for MERS-CoV

Figure 2. Propagation of Middle East respiratory syndrome coronavirus (MERS-CoV) in domestic pigs. We inoculated pigs intranasally and intratracheally with 106 tissue culture

infectious dose 50 (TCID50) of MERS- CoV isolate hCoV-EMC/2012 or, for controls (mock-inoculated), with Dulbecco’s modified Eagle medium. A) Mean bodyweight gain comparison between mock-inoculated and MERS- CoV–inoculated animals over time. Error bars indicate SDs. B) Mean viral loads shed from the nose and throat determined at the time points indicated by quantifying virus on nasal and throat swabs collected from MERS-CoV– inoculated animals using quantitative reverse transcription PCR (11); in each run, standard dilutions of a titered virus stock were run in parallel to calculate

TCID50 equivalents. Error bars indicate SDs. C) Viral loads in tissues collected from MERS-CoV–inoculated animals on days 4 (closed blue circles) and 11 (open red circles) postinoculation. Viral loads were determined as in panel B. Vertical bars indicate means. D) Serum samples collected from pigs at the time of euthanasia (days 4 and 11 postinoculation) and tested for MERS- CoV antibodies by using an ELISA for MERS-CoV spike 1 protein. Antibody titers are plotted as the reciprocal of the last serum dilution positive by ELISA. Horizontal bars indicate means. LN, lymph node.

Taken together, our data indicate that MERS-CoV can References infect pigs, leading to a low level of replication in the pig re- 1. World Health Organization. Middle East respiratory syndrome coronavirus (MERS-CoV)–Saudi Arabia. Disease outbreak news. spiratory tract, but does not cause clinical signs of disease. 2017 Mar 10 [cited 2017 Mar 15]. http://www.who.int/csr/don/ Furthermore, viral shedding from mucosal membranes of the 10-march-2017-mers-saudi-arabia/en/ upper respiratory tract was rather limited with no infectious 2. Ng DL, Al Hosani F, Keating MK, Gerber SI, Jones TL, virus measurable at any time postinoculation. Thus, the pig is Metcalfe MG, et al. Clinicopathologic, immunohistochemical, and ultrastructural findings of a fatal case of Middle East respiratory not a suitable animal disease model for MERS-CoV infection. syndrome coronavirus infection in the United Arab Emirates, April 2014. Am J Pathol. 2016;186:652–8. http://dx.doi.org/ Acknowledgments 10.1016/j.ajpath.2015.10.024 The authors thank the Rocky Mountain Veterinary Branch staff 3. Munster VJ, de Wit E, Feldmann H. Pneumonia from human coronavirus in a macaque model. N Engl J Med. 2013;368:1560–2. for the care and handling of the animals, Anita Mora for help http://dx.doi.org/10.1056/NEJMc1215691 with figure preparation, and Juergen Richt for helpful advice. 4. Baseler L, de Wit E, Feldmann H. A comparative review of animal models of Middle East respiratory syndrome coronavirus infection. Vet This work was supported by the Intramural Research Program of Pathol. 2016;53:521–31. http://dx.doi.org/ 10.1177/0300985815620845 the National Institute of Allergy and Infectious Diseases of the 5. Meurens F, Summerfield A, Nauwynck H, Saif L, Gerdts V. National Institutes of Health. The pig: a model for human infectious diseases. Trends Microbiol. 2012;20:50–7. http://dx.doi.org/10.1016/j.tim.2011.11.002 Dr. de Wit is a staff scientist in the Disease Modeling and 6. Müller MA, Raj VS, Muth D, Meyer B, Kallies S, Smits SL, et al. Transmission section of the Laboratory of Virology. Her Human coronavirus EMC does not require the SARS-coronavirus receptor and maintains broad replicative capability in mammalian research interests are in the pathogenesis and transmission of cell lines. MBio. 2012;3:e00515-12. http://dx.doi.org/10.1128/ emerging viruses. mBio.00515-12

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 987 DISPATCHES

7. van Doremalen N, Munster VJ. Animal models of Middle East 11. de Wit E, Rasmussen AL, Falzarano D, Bushmaker T, Feldmann F, respiratory syndrome coronavirus infection. Antiviral Res. Brining DL, et al. Middle East respiratory syndrome coronavirus 2015;122:28–38. http://dx.doi.org/10.1016/j.antiviral.2015.07.005 (MERS-CoV) causes transient lower respiratory tract infection in 8. Lu G, Hu Y, Wang Q, Qi J, Gao F, Li Y, et al. Molecular basis rhesus macaques. Proc Natl Acad Sci U S A. 2013;110:16598–603. of binding between novel human coronavirus MERS-CoV and http://dx.doi.org/10.1073/pnas.1310744110 its receptor CD26. Nature. 2013;500:227–31. http://dx.doi.org/ 12. Vergara-Alert J, van den Brand JM, Widagdo W, Muñoz M, 10.1038/nature12328 Raj S, Schipper D, et al. Livestock susceptibility to infection with 9. Wang N, Shi X, Jiang L, Zhang S, Wang D, Tong P, et al. Middle East respiratory syndrome coronavirus. Emerg Infect Dis. Structure of MERS-CoV spike receptor-binding domain 2017;23:232–40. http://dx.doi.org/10.3201/eid2302.161239 complexed with human receptor DPP4. Cell Res. 2013;23:986–93. http://dx.doi.org/10.1038/cr.2013.92 Address for correspondence: Heinz Feldmann, Laboratory of Virology, 10. Corman VM, Eckerle I, Bleicker T, Zaki A, Landt O, Division of Intramural Research, National Institute of Allergy and Infectious Eschbach-Bludau M, et al. Detection of a novel human coronavirus by real-time reverse-transcription polymerase chain reaction. Diseases, National Institutes of Health, Rocky Mountain Laboratories, 903 S Euro Surveill. 2012;17:20285. 4th St, Hamilton, MT 59840, USA; email: [email protected]

December 2011: Zoonotic Infections • Lineage and Virulence of • Seroprevalence of Streptococcus suis Alkhurma and Other Serotype 2 Isolates from Hemorrhagic Fever Viruses, North America Saudi Arabia • Isolation of Prion with • Knowledge of Avian BSE Properties from Influenza (H5N1) among Farmed Goat Poultry Workers, Hong • Candidate Cell Substrates, Kong, China • Fatal Outbreak of Vaccine Production, and • Risk for Human African Mycoplasma capricolum Pneumonia in Endangered • Risk for Rabies Importation Transmissible Spongiform Trypanosomiasis, Central from North Africa Encephalopathies Africa, 2000–2009 Markhors • Continuing Threat of • Molecular Epidemiology of • African Swine Fever Virus Influenza (H5N1) Virus Rift Valley Fever Virus Caucasus Isolate in Circulation in Egypt • Novel Multiplexed HIV/ European Wild Boars • Worldwide Occurrence Simian Immunodeficiency • Novel Sylvatic Rabies Virus and Impact of Human Virus Antibody Detection Variant in Endangered Trichinellosis, 1986–2009 Assay Golden Palm Civet, Sri Lanka • Sealpox Virus in Marine • Astroviruses in Rabbits Mammal Rehabilitation • Host Genetic Variants and • Rickettsia parkeri in Facilities, North America, Influenza-associated Amblyomma maculatum 2007–2009 Mortality among Children Ticks, North Carolina, • Transmission of Guanarito and Young Adults 2009–2010 and Pirital Viruses among • Severe Human Bocavirus • Japanese Encephalitis Virus Wild Rodents, Venezuela Infection, Germany • Animal Diseases Caused by Genotype Replacement, • Hepatitis E Virus in Rats, • Hepatitis E Virus Antibodies Orbiviruses, Algeria Taiwan, 2009–2010 Los Angeles, California in Blood Donors, France • Genogroup I and II • Altitude-dependent • Enterovirus Co-infections • Human Cardioviruses, Picobirnaviruses in Bartonella quintana and Onychomadesis after Meningitis, and Sudden Respiratory Tracts of Pigs Genotype C in Head Lice, Hand, Foot, and Mouth Infant Death Syndrome • Human Liver Infection by Ethiopia Disease, Spain in Children Amphimerus spp. Flukes, • Proximity to Goat Farms • Experimental Infection of Ecuador and Coxiella burnetii Horses with Hendra Virus/ • Q Fever in Woolsorters, Seroprevalence among Australia/Horse/2008/ Belgium Pregnant Women Redlands • Aedes aegyptiMosquitoes • Hemoptysis Associated • West Nile Virus Infection of Imported into the with Leptospirosis Acquired Birds, Mexico Netherlands, 2010 in Hawaii https://wwwnc.cdc.gov/eid/articles/issue/ 17/12/table-of-contents

988 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 High Rates of Neutralizing Antibodies to Toscana and Sandfly Fever Sicilian Viruses in Livestock, Kosovo

Nazli Ayhan, Kurtesh Sherifi,1 Arber Taraku, samples collected in the Pejë region contained TOSV Kristaq Bërxholi, Rémi N. Charrel1 neutralizing antibodies (6). Neutralizing antibody–based seroprevalence studies using animal serum proved inter- Toscana and sandfly fever Sicilian viruses (TOSV and esting regarding the global circulation of corresponding SFSV, respectively), both transmitted by sand flies, are viruses, as recently described in Portugal, Tunisia, and prominent human pathogens in the Old World. Of 1,086 Greece (7–9). To improve understanding of the circulation serum samples collected from cattle and sheep during 2013 in various regions of Kosovo (Balkan Peninsula), of TOSV and SFSV in Kosovo, we tested serum samples 4.7% and 53.4% had neutralizing antibodies against TOSV collected in cattle and sheep through neutralizing assay and and SFSV, respectively. field-trapped sand flies for viral RNA.

The Study hleboviruses (family Bunyaviridae, genus Phlebovi- In 2013, serum from domestic animals was collected Prus) are negative-sense tri-segmented RNA viruses from 12 different regions of Kosovo. Samples were col- for which mosquitoes, ticks, and sand flies are vectors. lected from 933 cattle and 153 sheep from 9 and 5 dif- In the Old World, phleboviruses transmitted by sand flies ferent regions, respectively, in Kosovo; the information (phlebotomines) are expanding in the Mediterranean ba- (location, specimen, date) were recorded. Ten milliliters sin, where an increasing number of new viruses have been of blood was taken from jugular venipuncture, and serum identified 1 ( ). There, sand fly–borne phleboviruses are was separated by centrifugation. All samples were stored divided into 3 groups in accordance with their antigenic at –20°C. relationships. Two groups correspond to recognized spe- We tested cattle and sheep serum for neutralizing cies: Sandfly fever Naples virus (including sandfly fever antibodies using the virus microneutralization assay, de- Naples [SFNV], Massilia, Tehran, and Toscana [TOSV] scribed for phleboviruses (7) in parallel with TOSV strain viruses) and Salehabad virus (including Salehabad and MRS2010-4319501 and SFSV strain Sabin. Serum sam- Arbia viruses). The third group comprises 2 viruses clas- ples were diluted from 1:10 to 1:80 into 96-well plates with sified as tentative species: Sandfly fever Sicilian virus a volume of 50 µL. Except for controls, we added a 1,000 (SFSV) and Corfou virus (2). Several are historic human 50% tissue culture infective dose in a 50-µL volume. For pathogens, such as SFSV and SFNV, which cause sandfly controls, we added 50 µL of Eagle’s minimum essential fever syndrome, a self-limited but severely incapacitating medium enriched with 5% fetal bovine serum, 1% peni- febrile illness (1); TOSV can cause central and periph- cillin–streptomycin, 1% L-glutamine 200 mmol/L, 1% ka- eral nervous system infections, such as meningitis and namycin, and 3% fungizone. The plates were incubated at encephalitis (3). 37°C. After 1 h, a 100 µL suspension of 2 × 105 Vero cells/ Although first data on sandfly fever were acquired mL was added and incubated at 37°C in the presence of from the Balkan region, few studies were published spe- 5% CO2. The microplates were read under an inverted mi- cifically about the situation in Kosovo (4): in 1976, a total croscope after 5 days for TOSV and 6 days for SFSV, and of 9.6% of human serum samples contained neutralizing the presence (neutralization titer at 10, 20, 40, and 80) or antibodies against SFSV (5) (the exact region of Kosovo absence (no neutralization) of cytopathic effect was noted. was not mentioned), and in 2011, <1% of human serum Cutoff value for positivity was set at titer> 20 (8). In 2014, a total of 267 sand flies were trapped and Author affiliations: Aix-Marseille University, Marseille, France identified. We tested these sand flies for phleboviruses us- (N. Ayhan, R.N. Charrel); Fondation Mediterranee Infection ing previously described protocols (9). Public Hospitals of Marseille, Marseille (N. Ayhan, R.N. Charrel); Global rates of TOSV neutralizing antibodies were in University of Hasan Prishtina, Prishtina, Kosovo (K. Sherifi); the same magnitude in cattle (5.14%) and sheep (1.96%) Agriculture University of Tirana, Triana, Albania (A. Taraku, in Kosovo (Table; Figure). Results observed in Pejë were K. Bërxholi) congruent with recent findings obtained with human

DOI: http://dx.doi.org/10.3201/eid2306.161929 1These authors contributed equally to this article.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 989 DISPATCHES

Table. Neutralizing antibodies against SFSV and TOSV in serum from cattle and sheep, and sandflies trapped in Kosovo, 2013* SFSV TOSV Serum source, region Total >20 Positive >20, % Total >20 Positive >20, % No. sand flies Sand fly species (%) Cattle, n = 933 546 58.5 48 5.1 Prizeren, n = 48 20 41.7 5 10.4 75 Phlebotomus major (70), P. simici (13), P. tobbi (8), P. papatasi (4), others (5) Pejë, n = 50 12 24.0 0 0 0 Rahovec, n = 198 101 51.0 15 7.6 2 P. major (100) Malishevë, n = 165 129 78.2 13 7.9 0 Glogovac, n = 50 35 70.0 0 0 0 Klinë, n = 50 39 78.0 5 10.0 0 Suharekë, n = 245 133 54.3 8 3.3 132 P. major (99), P. tobbi (1) Gjakovë, n = 50 35 70.0 2 4.0 0 Deçan, n = 77 42 54.6 0 0 0 Sheep, n = 153 34 22.2 3 2.0 Junik, n = 28 6 21.4 0 0 58 P. major (57), P. tobbi (38), others (5) Hani i Elezit, n = 50 7 14.0 2 4.0 0 Dragash, n = 30 17 56.7 1 3.3 0 Pejë, n = 8 3 37.5 0 0 0 Gjakovë, n = 37 1 2.7 0 0 0 Total, N = 1,086 580 53.4 51 4.7 267 *SFSV, sandfly fever Sicilian virus; TOSV, Toscana virus. serum; in both cases, TOSV circulation appears limited (6). antibodies in humans reported in the 1970s (5) most Although SFNV and TOSV belong to the same serocom- likely reflects circulation of SFNV rather than TOSV, a plex, they can be distinguished by using neutralization; finding that did not differ from our results and recent -re therefore, the high rate (27.9%) of SFNV neutralizing sults reported by others (6). Rates of TOSV neutralizing

Figure. Geographic distribution of rates of neutralizing antibodies against SFSV and TOSV in cattle and sheep, Kosovo, 2013. A) SFSV neutralizing antibodies in cattle. B) TOSV neutralizing antibodies in cattle. C) SFSV neutralizing antibodies in sheep. D) TOSV neutralizing antibodies in sheep. Inset in panel A shows location of Kosovo in Europe. SFSV, sandfly fever Sicilian virus; TOSV, Toscana virus.

990 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

TOSV and SFSV in Livestock, Kosovo antibodies were highest in southwestern regions of Koso- Conclusions vo, whereas negative results were obtained in the Pejë Our results confirm that TOSV and SFSV, or an SFSV-like area (Pejë, Deçan, Junik). Although we did not detect vi- virus, are circulating in several regions of Kosovo, which ral RNA in the 267 tested sand flies, TOSV was reported indicates that humans are exposed to these viruses. This in Croatia and Greece (10,), and a new phlebovirus was finding merits confirmation through seroprevalence studies described in Albania, Croatia, and Bosnia-Herzegovina and initiation of systematic testing for TOSV and SFSV (9; N. Ayhan et al., unpub. data). real-time reverse transcription PCR for febrile illness and Rates of SFSV neutralizing antibodies were much central nervous system infections during the warm season. higher than those for TOSV. Results for cattle ranged from 24.0% to 78.2% (mean 58.5%); results for sheep were low- This work was supported in part by the European Virus Archive er, ranging from 2.7% to 56.7% (mean 22.2%). For sheep, Goes Global project, which has received funding from the 4 of the 5 regions had rates of 14.0%–56.7%; the rate was European Union’s Horizon 2020 research and innovation much lower (2.7%) in Gjakovë. program under grant agreement no. 653316; the EDENext Few data are available for comparison; 9.6% of test- FP7-no. 261504 European Union project and this paper is ed human serum contained SFSV neutralizing antibodies catalogued by the EDENext Steering Committee as EDENext469 in the 1970s (5). Although no direct evidence (molecular (http://www.edenext.eu). The work of N.A. and R.N.C. was detection of viral RNA or virus isolation) exists of SFSV conducted under the frame of EurNegVec COST Action TD1303. or another SFSV-like virus in Kosovo or neighboring N.A. is supported by the Fondation Mediterranee Infection. countries, our results imply the presence of either SFSV Ms. Ayhan is a PhD student at Aix Marseille University. or an SFSV-related virus in Kosovo. We consider it valid Her research interests include phleboviruses transmitted by sand to use SFSV as a surrogate for all SFSV-related viruses flies in the Old World. (sandfly fever Turkey virus, sandfly fever Cyprus virus) because amino acid distances observed between the pro- References teins that elicit neutralizing antibodies (Gn and Gc) are 1. Alkan C, Bichaud L, de Lamballerie X, Alten B, Gould EA, well within the acceptable range, (i.e., <5% different for Charrel RN. Sandfly-borne phleboviruses of Eurasia and SFSV and SFSV-related viruses) (7). Thus, neutralizing Africa: epidemiology, genetic diversity, geographic range, antibodies are unlikely to discriminate between closely control measures. Antiviral Res. 2013;100:54–74. http://dx.doi.org/10.1016/j.antiviral.2013.07.005 related SFSV isolates. 2. Plyusnin A, Beaty BJ, Elliott RM, Goldbach R, Kormelink R, Recent seroprevalence studies showed high seroprev- Lundkvist A, et al. Bunyaviridae. In: King AMQ, Adams MJ, alence rates for SFSV neutralizing antibodies in dogs in Carstens EB, Lefkowitz EJ, editors. Virus taxonomy: classification Portugal (50.8%) (7), Tunisia (38.1%–59.2% depending on and nomenclature of viruses. Ninth report of the International Committee on Taxonomy of Viruses. San Diego (CA): Elsevier; the region) (8), and Greece (71.9%) and Cyprus (60.2%) 2011. p. 693–709. (11). Our results are congruent with data from continental 3. Charrel RN, Gallian P, Navarro-Mari JM, Nicoletti L, Papa A, Greece, with rates in the same order of magnitude (12). All Sánchez-Seco MP, et al. Emergence of Toscana virus in Europe. these findings verify the high prevalence of SFSV in the Emerg Infect Dis. 2005;11:1657–63. http://dx.doi.org/10.3201/ eid1111.050869 Mediterranean basin. Because of the nature of this study, 4. Pick A. Zur pathologie und therapie einer eigenthumlichen the serum was collected and stored under conditions that endemischen Krankheitsform. Wien Med Wochenschr. 1886; prevented attempts to detect viral RNA and isolate viral 33:1141–5. strains. SFSV remains an important human pathogen, as 5. Tesh RB, Saidi S, Gajdamovič SJA, Rodhain F, Vesenjak-Hirjan J. Serological studies on the epidemiology of sandfly fever in the Old recently highlighted in Africa and Turkey (12,13). World. Bull World Health Organ. 1976;54:663–74. Published data about the distribution of sand flies and 6. Venturi G, Marchi A, Fiorentini C, Ramadani N, Quaglio G, species identification are old and scarce; Phlebotomus pa- Kalaveshi A, et al. Prevalence of antibodies to phleboviruses and patasi, P. perfiliewi, P. neglectus, and P. tobbi sand flies flaviviruses in Peja, Kosovo. Clin Microbiol Infect. 2011;17:1180– 2. http://dx.doi.org/10.1111/j.1469-0691.2010.03445.x have been documented in Kosovo and neighboring coun- 7. Alwassouf S, Maia C, Ayhan N, Coimbra M, Cristovao JM, tries with similar environmental and climatic conditions Richet H, et al. Neutralization-based seroprevalence of Toscana (14,15). In our study, P. major sand flies dominated (81%), virus and sandfly fever Sicilian virus in dogs and cats from followed by P. tobbi (11.2%), P. simici (3.7%), P. papatasi Portugal. J Gen Virol. 2016;97:2816–23. http://dx.doi.org/10.1099/ jgv.0.000592 (1%), and other species (3.1%). 8. Sakhria S, Alwassouf S, Fares W, Bichaud L, Dachraoui K, SFSV positivity varied among the regions for cattle Alkan C, et al. Presence of sandfly-borne phleboviruses of two and sheep. The regional prevalence differences might have antigenic complexes (sandfly fever Naples virus and sandfly fever resulted from geographic and climatic characteristics of the Sicilian virus) in two different bio-geographical regions of Tunisia demonstrated by a microneutralisation-based seroprevalence study region that could affect the vector sand fly species distribu- in dogs. Parasit Vectors. 2014;7:476. http://dx.doi.org/10.1186/ tion and population size. s13071-014-0476-8

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 991 DISPATCHES

9. Ayhan N, Velo E, de Lamballerie X, Kota M, Kadriaj P, Ozbel Y, 13. Woyessa AB, Omballa V, Wang D, Lambert A, Waiboci L, Ayele W, et al. Detection of Leishmania infantum and a novel phlebovirus et al. An outbreak of acute febrile illness caused by sandfly fever (Balkan virus) from sand flies in Albania. Vector Borne Zoonotic Sicilian virus in the Afar region of Ethiopia, 2011. Am J Trop Med Dis. 2016;16:802–6. http://dx.doi.org/10.1089/vbz.2016.2002 Hyg. 2014;91:1250–3. http://dx.doi.org/10.4269/ajtmh.14-0299 10. Papa A, Paraforou T, Papakonstantinou I, Pagdatoglou K, 14. Simitch T, Zivkovitch V. Phlebotome fauna in Yugoslavia and Kontana A, Koukoubani T. Severe encephalitis caused by their role in the epidemiology of pappataci fever, kala-azar and Toscana virus, Greece. Emerg Infect Dis. 2014;20:1417–9. cutaneous leishmaniasis [in French]. Arch Inst Pasteur Alger. http://dx.doi.org/10.3201/eid2008.140248 1956;34:380–7. 11. Alwassouf S, Christodoulou V, Bichaud L, Ntais P, Mazeris A, 15. Živković V. Faunistic and ecological investigation of sandflies Antoniou M, et al. Seroprevalence of sandfly-borne phleboviruses (Diptera, Phlebotomidae) in Serbia. Acta Vet (Beogr). 1980; belonging to three serocomplexes (sandfly fever Naples, sandfly 30:67–88. fever Sicilian and Salehabad) in dogs from Greece and Cyprus using neutralization test. PLoS Negl Trop Dis. 2016;10:e0005063. Address for correspondence: Rémi N. Charrel, Aix Marseille http://dx.doi.org/10.1371/journal.pntd.0005063 University, UMR_D 190 Emergence des Pathologies Virales, Faculty 12. Ergunay K, Ayhan N, Charrel RN. Novel and emergent sandfly- borne phleboviruses in Asia Minor: a systematic review. Rev Med of Medicine, 27 blvd Jean Moulin, Marseille 13385, France; email: Virol. 2017; 27:e1898. http://dx.doi.org/10.1002/rmv.1898 [email protected] May 2016: Vectorborne Diseases • Expansion of Shiga Toxin– Producing Escherichia coli by Use of Bovine Antibiotic Growth Promoters • Projecting Month of Birth for At-Risk Infants after Zika Virus Disease Outbreaks • Genetic Characterization of Archived Bunyaviruses and Their Potential for Emergence • An Operational Framework in Australia for Insecticide Resistance • Rickettsia parkeri Rickettsiosis, Management Planning Arizona, USA • Plasmodium falciparum K76T pfcrt Gene Mutations and • Plasmodium falciparum Parasite Population Structure, In Vitro Resistance to Haiti, 2006–2009 Monodesethylamodiaquine, • Outbreak of Middle East Dakar, Senegal, 2014 Respiratory Syndrome at • Astrovirus MLB2, a New Tertiary Care Hospital, Jeddah, Gastroenteric Virus Associated Saudi Arabia, 2014 with Meningitis and • Differences in Genotype, Disseminated Infection Clinical Features, and • Spectrum of Viral Pathogens Inflammatory Potential of in Blood of Malaria-Free Ill Borrelia burgdorferi sensu Travelers Returning to Canada stricto Strains from Europe and and the United States • Expanded Geographic Distribution and Clinical • Acute Human Inkoo and Characteristics of Ehrlichia Chatanga Virus Infections, ewingii Infections, Finland United States

https://wwwnc.cdc.gov/eid/articles/ issue/22/5/table-of-contents

992 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Congenital Malformations of Calves Infected with Shamonda Virus, Southern Japan

Yoshimasa Hirashima, Shoei Kitahara, no detailed description of an association between SHAV Tomoko Kato, Hiroaki Shirafuji, and malformations, we report details of these 15 clinical Shogo Tanaka, Tohru Yanase cases of malformations in calves suspected to be caused by SHAV infection. In 2015 and 2016, we observed 15 malformed calves that were exposed to intrauterine infection with Shamonda virus, The Study a Simbu serogroup orthobunyavirus, in Japan. Characteristic To obtain data on arboviruses circulating in 2015, we manifestations were arthrogryposis and gross lesions in the central nervous system. Our results indicate that this arbo- attempted to isolate viruses on BHK-21 and HmLu-1 virus should be considered a teratogenic virus in ruminants. cells inoculated with blood samples obtained from 60 sentinel cattle maintained on 30 farms and from pools of Culicoides biting midges collected by using suction he Simbu virus serogroup is composed of >25 sero- light traps on 2 cattle farms in Kagoshima Prefecture in Tlogically related viruses in the family Bunyaviridae, southern Japan. Two viruses (KS-1/P/15 and KS-2/P/15) genus Orthobunyavirus (1), which are transmitted mainly were isolated from cattle blood collected during August by Culicoides biting midges. Several of these viruses, such and September 2015, and another virus (KSB-1/C/15) as Akabane virus, Aino virus, and Schmallenberg virus, are was isolated from a pool of C. tainanus midges sampled arboviruses associated with abortion, premature birth, still- during September 2015. birth, and congenital malformations in ruminants (2–4). We performed reverse transcription PCR (RT-PCR) The emergence and spread of Schmallenberg virus with primer pairs (AKAI206F; 5′-CACAACCAAgTgTC- has had large socioeconomic effects in countries in Europe gATCTTA-3′; and SimbuS637–656; 5′-gAgAATCCA- (4,5). Frequent epizootics of Akabane virus and Aino virus gATTTAgCCCA-3′) specific for small RNA segment of in Japan have caused many cases of congenital malforma- Simbu serogroup viruses and the One Step RT-PCR Kit tions in calves (6). However, the etiologic diagnosis for (QIAGEN, Hilden, Germany). We generated a product of malformed calves associated with other arboviruses has not the expected size from RNA samples of the isolated virus- been established because of a lack of knowledge and sensi- es. Preliminary sequence analysis for the RT-PCR prod- tive diagnostic systems. Attempts to isolate viruses from uct (443-nt) showed that the viruses were highly similar to sentinel cattle and Culicoides biting midges have contrib- SHAV. We sequenced and analyzed complete small and uted to knowledge about arboviruses circulating in nature medium RNA segments and a partial region of the large and have, in some instances, helped predict the etiologic RNA segment by using primers specific for SHAV (8). Se- agents responsible for malformations (7). quences determined in this study were deposited in the In- Three Simbu serogroup viruses, Peaton virus, Sathu- ternational Nucleotide Sequence Database under accession peri virus, and Shamonda virus (SHAV), were identified nos. LC198185–93. in Japan during the past 2 decades and have been suspect- Neighbor-joining analysis available in MEGA7 (9) ed of being involved in congenital defects in calves (8). was used for phylogenetic analysis on the basis of the 3 During December 2015–April 2016 in southern Japan, RNA segments of the Simbu serogroup viruses. Sequences SHAV infections were identified in 15 malformed calves determined showed high nucleotide identities with known that had no antibodies against other teratogenic viruses. Of sequences of SHAV (98.3%–99.5% for the RNA small the 3 segments of the RNA genome of SHAV, the small segment, 89.0%–97.9% for the medium RNA segment, and and large segments have high genetic similarity with those 91.5%–98.0% for the large RNA segment). Three phyloge- of Schmallenberg virus, which implies the teratogenic- netic trees showed that isolated viruses densely clustered ity of SHAV in the ruminant fetus (8). Because there is with Japanese SHAV isolates obtained in 2002 and 2007 Author affiliations: Kagoshima Central Livestock Hygiene Service (Figure 1). Center, Hioki, Japan (Y. Hirashima, S. Kitahara); National We performed virus neutralization tests (VNTs) on Agriculture and Food Research Organization, Kagoshima, Japan virus-infected HmLu-1 cells by using an established meth- (T. Kato, H. Shirafuji, S. Tanaka, T. Yanase) od (2). Antibodies to SHAV (titer range 1:2–1:64) were detected in serum samples from 15 malformed calves by DOI: http://dx.doi.org/10.3201/eid2306.161946 VNTs during December 2015–April 2016 (Table). Serum

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 993 DISPATCHES

samples obtained from sentinel cattle in June, August, Sep- tember, October, and November 2015 showed that seroconversion for SHAV had occurred widely in Kagoshima Prefecture during August–October. Although 2 calves ingested colostrum substitute containing immunoglobulins, all calves tested were deprived of colostrum produced by their mothers. SHAV had not been detected in mainland Japan for ≈10 years until we identified new cases of infection in 2015. It is highly improbable that the colostrum substitute contained antibodies against SHAV. Malformed calves were delivered at or beyond term (gestation periods range 275–299 days), but 6 calves were born dead. The small RNA–specific RT-PCR showed positive results for samples from the central nervous system (cerebrum, brainstem, or spinal cord) of 7 deformed calves. Se- quences obtained were identical (except for 1 of 2 nt substitutions) with relevant sequence of SHAVs isolated in 2015. Torticollis (10/15) or arthrogryposis (12/15) were often observed among affected calves (Figure 2). Head deformity (6/15) or spinal curvature (11/15) were also characteristic. Discoloration of skeletal muscles was observed in one third of the calves. Cerebral hemispheres appeared normal among the calves, but lateral ventricular enlargement was found in 2 calves, and cerebellar hypoplasia was found in 1 calf. Histopathologic analysis identified calcification of nerve cells (11/15), perivascular infiltration of mononuclear cells (6/15), and gliosis (5/15), which were often observed in the cerebrum and brainstem (Figure 2). Severe degen- erative changes in the ventral horn of the spinal cord were identified in 10 of the malformed calves. Fatty replacement (12/15), atrophy (9/15), and myositis (9/15) were major observations in the skeletal muscle and often correlated with muscle discolorations. VNTs did not detect neutralizing antibodies against teratogenic arboviruses, such as Chuzan virus, Akabane virus, Aino virus, or Peaton virus in serum samples from affected calves. All tested dams of affected newborns were positive for SHAV antibodies in VNTs (titer range 1:16–>1:256).

Conclusions Our findings support an association between SHAV and congenital deformities in calves infected in utero. Mani- festations and macroscopic and microscopic observations in the malformed calves were similar to those of calves infected with Schmallenberg virus (10,11). In compari- Figure 1. Neighbor-joining phylogenetic trees based on protein- son with cases attributed to Akabane virus and Aino virus, coding sequences of A) small, B) medium, and C) large partial congenital lesions in the brain were relatively mild (i.e., no RNA segments for Simbu serogroup viruses, southern Japan, hydranencephaly and few lateral ventricular enlargements 2015–2016. Black circles indicate Shamonda viruses isolated appeared in malformed calves). However, nonsuppurative in this study. Values along branches are percentages (≥50%) of encephalitis or nerve cell death often occurred in cerebrums bootstrap support of 1,000 pseudoreplicates. The 3 segmented RNAs of Bunyamwera virus were used as outgroups to root the and brainstems that otherwise seemed normal. trees. Scale bars indicate nucleotide substitutions per site. Twelve of the dams of the malformed calves in this NA, details not available. report were vaccinated with a trivalent inactivated vaccine

994 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Malformations of Calves with Shamonda Virus

Table. Characteristics of 15 malformed calves infected with Shamonda virus, southern Japan, December 2015 ̶ April 2016* Calf no. Characteristic 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Total Gestational age, d 281 275 280 278 285 291 281 280 293 279 290 287 287 276 299 NA Euthanasia + + + – + + + + + – – – + – – 9 Stillbirth – – – + – – – – – + + + – + + 6 RT-PCR result + – – + + + – – – + – + – – + 7 Antibody titer 1:8 1:32 1:32 1:64 1:2 1:32 1:32 1:16 1:8 1:16 1:4 1:16 1:16 1:4 1:64 NA Clinical finding Torticollis + – + + + – + + – + – + – + + 10 Arthrogryposis – – – + + + + + + + + + + + + 12 Macroscopic finding Head deformity Brachygnathism + + + – – – – – – – – – + – – 4 Asymmetry of skull + – + – – – – – – – + + – – – 4 LVE – – – – – – – – – + – + – – – 2 Cerebellar hypoplasia – – – – – – + – – – – – – – – 1 Spinal curvature + – + + + – + + – + + + – + + 11 Muscle discoloration – – – – + + – – + + + – – – – 5 Histopathologic finding Cerebrum Calcification of nerve cells – + + – + – + – – – – + – – – 5 Brainstem Calcification of nerve cells – – + + – + – + – + – + – + + 8 Perivascular infiltration – + – – – + + + + + – – – – – 6 Gliosis – – + + – + – – – + – – – + – 5 Spinal cord Decrease/disappearance – – – + – + + – + + + + + + + 10 of ventral horn cells Skeletal muscles Fatty replacement + – + – + + + – + + + + + + + 12 Atrophy + – + – + + + – + – + – – + + 9 Myositis + – + – + + + – + – + – – + + 9 *LVE, lateral ventricular enlargement; NA, not applicable; RT-PCR, reverse transcription PCR; +, positive; –, negative. containing Akabane, Aino, and Chuzan viruses. To our previously unaffected areas, because long-distance dispersal knowledge, no effective preventive measure for infection and accidental transportation of infected vectors from epizo- with SHAV is available. Previous surveillance in Africa, the otic areas can introduce the virus. Also, recent outbreaks of Middle East, and Asia (12–14) enabled us to postulate the infection with Schmallenberg virus and SHAV suggest that wide geographic distribution of SHAV. The potential risk for many Simbu serogroup viruses can affect livestock. More SHAV spreading in livestock should be considered, even in detailed study of this virus serogroup is warranted.

Figure 2. Characteristic observations in Shamonda virus–positive malformed calves, southern Japan, 2015–2016. A) Torticollis and arthrogryposis in calf 3. B) Spinal curvature (scoliosis) in calf 7. C) Perivascular infiltration in the midbrain of calf 7. D) Fatty replacement and atrophy in skeletal muscle of calf 3. For histopathologic analysis, thin sections prepared from paraffin-embedded tissues were stained with hematoxylin and eosin. Scale bars indicate 50 µm.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 995

DISPATCHES

Acknowledgments cattle in southern Japan from 2003 to 2013. Transbound We thank veterinary officers at Kagoshima Prefectural livestock Emerg Dis. 2016;63:e160–72. http://dx.doi.org/10.1111/tbed.12324 8. Yanase T, Kato T, Aizawa M, Shuto Y, Shirafuji H, Yamakawa M, hygiene service centers and technical staff at Kyushu Research et al. Genetic reassortment between Sathuperi and Shamonda Station, National Institute of Animal Health, for excellent techni- viruses of the genus Orthobunyavirus in nature: implications cal support and sampling. for their genetic relationship to Schmallenberg virus. Arch Virol. 2012;157:1611–6. http://dx.doi.org/10.1007 Mr. Hirashima is a veterinary officer at the Kagoshima Central /s00705-012-1341-8 Livestock Hygiene Service Center, Hioki, Japan. His research 9. Kumar S, Stecher G, Tamura K. MEGA7: Molecular Evolutionary interests include ruminant arboviruses associated with abnormal Genetics Analysis version 7.0 for bigger datasets. Mol Biol Evol. 2016;33:1870–4. http://dx.doi.org/10.1093/molbev/msw054 births and febrile diseases in southern Japan. 10. Bayrou C, Garigliany MM, Sarlet M, Sartelet A, Cassart D, Desmecht D. Natural intrauterine infection with Schmallenberg virus in malformed newborn calves. Emerg Infect Dis. References 2014;20:1327–30. http://dx.doi.org/10.3201/eid2008.121890 1. Elliott RM, Blakqori G. Molecular biology of orthobunyaviruses. In: 11. Peperkamp NH, Luttikholt SJ, Dijkman R, Vos JH, Junker K, Plyusnin A, Elliott RM, editors. Bunyaviridae: molecular and cellular Greijdanus S, et al. Ovine and bovine congenital abnormalities biology. Norfolk (UK): Caister Academic Press; 2011. p. 1–39. associated with intrauterine infection with Schmallenberg 2. Tsuda T, Yoshida K, Ohashi S, Yanase T, Sueyoshi M, Kamimura S, virus. Vet Pathol. 2015;52:1057–66. http://dx.doi.org/10.1177/ et al. Arthrogryposis, hydranencephaly and cerebellar hypoplasia 0300985814560231 syndrome in neonatal calves resulting from intrauterine infection 12. Yanase T, Maeda K, Kato T, Nyuta S, Kamata H, Yamakawa M, with Aino virus. Vet Res. 2004;35:531–8. http://dx.doi.org/10.1051/ et al. The resurgence of Shamonda virus, an African Simbu vetres:2004029 group virus of the genus Orthobunyavirus, in Japan. Arch Virol. 3. Kirkland PD. Akabane virus infection. Rev Sci Tech. 2015;34:403– 2005;150:361–9. http://dx.doi.org/10.1007/s00705-004-0419-3 10. http://dx.doi.org/10.20506/rst.34.2.2366 13. Causey OR, Kemp GE, Causey CE, Lee VH. Isolations of 4. Afonso A, Abrahantes JC, Conraths F, Veldhuis A, Elbers A, Simbu-group viruses in Ibadan, Nigeria 1964-69, including the Roberts H, et al. The Schmallenberg virus epidemic in new types Sango, Shamonda, Sabo and Shuni. Ann Trop Med Europe2011–2013. Prev Vet Med. 2014;116:391–403. Parasitol. 1972;66:357–62. http://dx.doi.org/10.1080/00034983.19 http://dx.doi.org/10.1016/j.prevetmed.2014.02.012 72.11686835 5. Hoffmann B, Scheuch M, Höper D, Jungblut R, Holsteg M, 14. Brenner J. Preliminary report of serosurvey for circulation of Schirrmeier H, et al. Novel orthobunyavirus in cattle, Europe, viruses of the Simbu serogroup in dairy cattle in Israel, 2014. 2011. Emerg Infect Dis. 2012;18:469–72. http://dx.doi.org/10.3201/ ProMed. 2016 Apr 17 [cited 2016 Dec 1]. http://www.promedmail. eid1803.111905 org, archive no. 20160417.4165684. 6. Forman S, Hungerford N, Yamakawa M, Yanase T, Tsai HJ, Joo YS, et al. Climate change impacts and risks for animal health Address for correspondence: Tohru Yanase, Kyushu Research Station, in Asia. Rev Sci Tech. 2008;27:581–97. http://dx.doi.org/10.20506/ rst.27.2.1814 National Institute of Animal Health, National Agriculture and Food 7. Kato T, Shirafuji H, Tanaka S, Sato M, Yamakawa M, Tsuda T, Research Organization, 2702 Chuzan, Kagoshima 891-0105, Japan; et al. Bovine arboviruses in Culicoides biting midges and sentinel email: [email protected]

EID Adds Advanced Search Features for Articles Emerging Infectious Diseases now has an advanced search feature that makes it easier to find articles by using keywords, names of authors, and specified date ranges. You can sort and refine search results by manuscript number, volume or issue number, or article type. A quick start guide and expandable help section show you how to optimize your searches. https://wwwnc.cdc.gov/eid/AdvancedSearch EID’s new mapping feature allows you to search for articles from specific countries by using a map or table to locate countries. You can refine search results by article type, volume and issue, and date, and bookmark your search results. https://wwwnc.cdc.gov/eid/ArticleMap

996 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Brucella neotomae Infection in Humans, Costa Rica

Marcela Suárez-Esquivel, Rica. This isolate (denoted strain bneohCR2) was cultured Nazareth Ruiz-Villalobos, César Jiménez-Rojas, from a cerebrospinal fluid sample obtained from a 64-year- Elías Barquero-Calvo, Carlos Chacón-Díaz, old male patient at one of the main hospitals in San José, Eunice Víquez-Ruiz, Norman Rojas-Campos, Costa Rica. In 2011, another isolate (denoted strain bneo- Kate S. Baker,1 Gerardo Oviedo-Sánchez, hCR1) was recovered from a cerebrospinal fluid sample Ernesto Amuy, Esteban Chaves-Olarte, from a 51-year-old male patient at a regional hospital in Nicholas R. Thomson, Edgardo Moreno, Costa Rica. As is common for other patients with brucel- Caterina Guzmán-Verri losis, the blood leukocyte count for each patient was almost within the reference range, and C-reactive protein level was Several species of Brucella are known to be zoonotic, but within reference range. Both patients showed clinical signs B. neotomae infection has been thought to be limited to compatible with neurobrucellosis (6), had positive Rose wood rats. In 2008 and 2011, however, B. neotomae was Bengal test results, and recovered after receiving 1 month isolated from cerebrospinal fluid of 2 men with neurobrucellosis. The nonzoonotic status of B. neotomae should of streptomycin (750 mg/d) and 3 months of doxycycline be reassessed. (100 mg/12 h). Further bacteriologic analysis (7,8) confirmed that the isolates were a Brucella sp. (Table). Real-time PCR high- embers of the genus Brucella are the infectious resolution melting analysis (9) confirmed genus designa- Magents of brucellosis, a neglected disease respon- tion but was inconclusive regarding species designation. sible for economic losses resulting from abortion and low Bruce-ladder multiplex PCR (10) and multiple loci variable performance in production animals (1). The 4 species number of tandem repeats–16-loci panel analysis (http:// mainly responsible for this widespread bacterial zoono- mlva.u-psud.fr/brucella/; Figure 1) indicated that the pro- sis are B. melitensis, B. abortus, B. suis, and to a lesser files for both DNA isolates corresponded to profiles for B. extent B. canis. Underdiagnosis and limited awareness of neotomae. Analysis of bneohCR2 by multiplex single-nu- the disease among healthcare practitioners is common in cleotide polymorphism (SNP) primer extension assay (11) many countries (1). and by matrix-assisted laser desorption/ionization time-of- B. neotomae, isolated in 1957 from wood rats (Neo- flight mass spectrometry of protein extracts (12) (online toma lepida) in North America (2), has been considered Technical Appendix Table 1, https://wwwnc.cdc.gov/EID/ nonzoonotic (3). It has been isolated from target organs of article/23/6/16-2018-Techapp1.pdf) confirmed that the iso- experimentally infected mice and guinea pigs (4,5). We re- late was B. neotomae. port the isolation of B. neotomae from cerebrospinal fluid We performed whole-genome sequencing of bneohCR1 samples from 2 human patients with neurobrucellosis. and bneohCR2 and resequencing of reference strain B. neotomae 5K33. Sequencing data were deposited at the Euro- The Study pean Nucleotide Archive (http://www.ebi.ac.uk/ena/) under In 2008, a Brucella sp. isolate was submitted to the Tropi- accession codes ERS1563929 (bneohCR1), ERS1563928 cal Diseases Research Center at the Universidad de Costa (bneohCR2), and ERS1624467 (SK33). To place the bneo- hCR1 and bneohCR2 in a phylogenetic context, publicly Author affiliations: Universidad Nacional, Heredia, Costa Rica available reads from 51 Brucella whole-genome sequences (M. Suárez-Esquivel, N. Ruiz-Villalobos, C. Jiménez-Rojas, (online Technical Appendix Table 2) were aligned and then E. Barquero-Calvo, E. Víquez-Ruiz, E. Moreno, C. Guzmán-Verri); mapped to B. suis 1330 by using SMALT version 0.5.8 (ftp:// Universidad de Costa Rica, San José, Costa Rica ftp.sanger.ac.uk/pub/resources/software/smalt/). Reads from (E. Barquero-Calvo, C. Chacón-Díaz, N. Rojas-Campos, G. bneohCR1 and bneohCR2 genomes mapped to 98.6% of Oviedo-Sánchez, E. Chaves-Olarte, E. Moreno, C. Guzmán-Verri); the B. suis 1330 genome. SNPs were called from the align- Wellcome Trust Sanger Institute, Hinxton, UK (K.S. Baker, ment by use of Samtools (http://samtools.sourceforge.net/), N.R. Thomson); Caja Costarricense del Seguro Social, and 34,307 variable sites across all isolates were extracted Puntarenas, Costa Rica (E. Amuy) by using SNP sites (13). The resulting alignment of SNPs DOI: http://dx.doi.org/10.3201/eid2306.162018 1Current affiliation: University of Liverpool, Liverpool, UK.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 997 DISPATCHES

Table. Differential biochemical profile of Brucella isolates from 2 (https://github.com/sanger-pathogens/Farm_blast) compari- men with neurobrucellosis, Costa Rica, 2008 and 2011 son of this pseudo-molecule and the bneohCR draft genomes, Analysis bneohCR1 bneohCR2 generated by assembly with Velvet (15), showed that the ge- Biochemical tests nomic loci known as 26.5 kb, 12 kb, and GI-6 that are absent Oxidase - – Citrate utilization – – in B. neotomae (14) are also absent in the queried genomes. Nitrate reduction + + CO2 required – – Conclusions H2S production + + Urease activity, h 0–0.5+ 0–0.5+ This report of B. neotomae as a cause of zoonotic disease Growth in presence of dyes raises questions about possible underrepresentation of re- Thionin 20 g/mL – – ported cases. This study also has implications for brucellosis Basic fuchsin 20 g/mL – – diagnosis. Specifically, the differences among B. neotomae A + + M – – and the other Brucella species at the biochemical level are subtle. The major difference between B. neotomae and B. was used for maximum-likelihood phylogenetic reconstruc- abortus, the main cause of human brucellosis in Costa Rica, tion by use of RAxML version 7.0.4 (https://github.com/ is the presence of oxidase activity in B. abortus, which is stamatak/standard-RAxML). The generated phylogenetic assessed subjectively (7,8). Because B. neotomae has not, tree confirmed that the bneohCR isolates clustered together until now, been considered zoonotic, some cases of brucel- with B. neotomae reference strain 5K33 (ENA accession no. losis reported as being caused by atypical zoonotic classical JMSC01, assembly accession no. GCA_00742255.1) (Fig- Brucella might have been misdiagnosed cases of B. neoto- ure 2). Isolates bneohCR1 and bneohCR2 differed from the mae infection. The introduction of whole-genome sequenc- reference genome by 174 and 160 SNPs, respectively. This ing into the clinical field will thus improve diagnosis. number of SNPs is smaller than that between B. abortus A lack of epidemiologic information with regard to the 9–941 and B. abortus 2308 (214 SNPs), which are 2 well- 2 patients reported here precluded the investigation of ex- recognized strains of the same Brucella species. posure or contact with hosts known to harbor B. neotomae. Analysis of 23 previously reported genomic islands or Further studies are needed to establish which animals may anomalous genomic loci (14) was performed for both bneo- act as reservoirs for B. neotomae in Costa Rica. hCR genomes. For this analysis, a “genomic-island pseudo- In summary, B. neotomae should be considered a molecule” was constructed by concatenation of 23 genomic zoonosis risk for infection in humans. Incorporation of regions obtained from different Brucella genomes. BLAST molecular techniques for diagnosis will help resolve the

Figure 1. Dendrogram based on multiple locus variable number of tandem repeats–16-loci panel analysis of Brucella spp. (http://mlva.u-psud.fr/brucella/) and clinical isolates from human cerebrospinal fluid samples from 2 patients with brucellosis. The isolates bneohCR1 and bneohCR2 (red branches) showed a pattern consistent with previously reported profiles for Brucella neotomae (blue shading). Black, gray, and tic marks are used to differentiate between adjacent species. Arrows separate small neighboring clusters and indicate the B. neotomae cluster.

998 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Brucella neotomae Infection in Humans

Figure 2. Phylogenetic tree based on 34,307 single-nucleotide polymorphisms (SNPs) found among 51 Brucella genome sequences. The clinical isolates bneohCR1 and bneohCR2 cluster with B. neotomae 5K33 and differ by 164 SNPs. A different color is used to represent each Brucella species. Dotted red lines denote the 3 B. neotomae isolates, which overlap at the tip of the branch because of the high identity among them.

Brucella genus homogeneity obtained when only biochem- References ical assays are used. 1. World Health Organization. The control of neglected zoonotic diseases. In: NZD4 Organising Committee, editor. WHO conference report. Geneva: WHO Press; 2014. p. 23–35. Acknowledgments 2. Stoenner HG, Lackman DB. A new species of Brucella isolated We are grateful to Heilyn Estrada-Murillo for isolate from the desert wood rat, Neotoma lepida Thomas. Am J Vet Res. information, Daphnne Garita for technical assistance, Gabriela 1957;18:947–51. 3. Moreno E. Retrospective and prospective perspectives on zoonotic Hernandez-Mora for provision of serotype-specific antiserum, brucellosis. Front Microbiol. 2014;5:213. http://dx.doi.org/10.3389/ and Gordon Dougan for helpful discussions. fmicb.2014.00213 4. Stoenner HG. The behavior of Brucella neotomae and Brucella suis This work was funded by Fondos del Sistema del Consejo in reciprocal superinfection experiments in mice and guinea pigs. Nacional de Rectores, Costa Rica, and Wellcome Trust. N.R.-V. Am J Vet Res. 1963;24:376–80. and C.J.-R. were partially sponsored by University of Costa Rica 5. Gibby IW, Gibby AM. Host–parasite relationships with Brucella scholarships. K.S.B. is funded by Wellcome Trust Fellowship neotomae. J Bacteriol. 1965;89:9–16. 6. Ceran N, Turkoglu R, Erdem I, Inan A, Engin D, Tireli H, et al. 106690/A/14/Z. K.S.B. and N.R.T. were funded by Wellcome Neurobrucellosis: clinical, diagnostic, therapeutic features and Trust grant no. 098051. outcome. Unusual clinical presentations in an endemic region. Braz J Infect Dis. 2011;15:52–9. http://dx.doi.org/10.1016/ All procedures involving live Brucella were carried out according S1413-8670(11)70140-4 to the “Reglamento de Bioseguridad de la CCSS39975-0,” year 7. Alton GG, Jones LM, Angus RD, Verger JM. Techniques for the 2012, after “Decreto Ejecutivo no. 0965-S,” year 2002 and brucellosis laboratory. Paris: Institut National de la Recherche Agronomique; 1988. p. 39–41. protocol approved by SIA0434-14 Universidad Nacional, Costa 8. Weyant R, Popovic T, Bragg S. Basic laboratory protocols for the Rica. These genetic resources were accessed in Costa Rica presumptive identification ofBrucella species. Atlanta: Centers for according to the Biodiversity Law no. 7788 and the Convention on Disease Control and Prevention; 2001. p. 1–16. Biological Diversity, under the terms of respect to equal and fair 9. Winchell JM, Wolff BJ, Tiller R, Bowen MD, Hoffmaster AR. Rapid identification and discrimination ofBrucella isolates by distribution of benefits among those who provided such resources. use of real-time PCR and high-resolution melt analysis. Miss Suárez-Esquivel is a PhD student at Programa de J Clin Microbiol. 2010;48:697–702. http://dx.doi.org/10.1128/ JCM.02021-09 Doctorado en Ciencias as Universidad de Costa Rica, working at 10. García-Yoldi D, Marín CM, de Miguel M-J, Muñoz PM, Programa de Investigación en Enfermedades Tropicales at Vizmanos JL, López-Goñi I. Multiplex PCR assay for the Universidad Nacional. Her primary research interest is the identification and differentiation of allBrucella species and the evolution of Brucella species and their host specificity. vaccine strains Brucella abortus S19 and RB51 and Brucella

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 999 DISPATCHES

melitensis Rev1. Clin Chem. 2006;52:779–81. http://dx.doi.org/ 14. Mancilla M. The Brucella genomic islands. In: López-Goñi I, 10.1373/clinchem.2005.062596 O’Callaghan D, editors. Brucella: Molecular Microbiology 11. Scott JC, Koylass MS, Stubberfield MR,Whatmore AM. and Genomics. Poole (UK): Caister Academic Press; 2012. Multiplex assay based on single-nucleotide polymorphisms for p. 36–57. rapid identification ofBrucella isolates at the species level. Appl 15. Page AJ, De Silva N, Hunt M, Quail MA, Parkhill J, Harris SR, Environ Microbiol. 2007;73:7331–7. http://dx.doi.org/10.1128/ et al. Robust high-throughput prokaryote de novo assembly AEM.00976-07 and improvement pipeline for Illumina data. Microb Genomics. 12. Isidoro-Ayza M, Ruiz-Villalobos N, Pérez L, Guzmán-Verri C, Microbiology Society; 2016 Aug 25. http://dx.doi.org/10.1099/ Muñoz PM, Alegre F, et al. Brucella ceti infection in dolphins from mgen.0.000083 the western Mediterranean Sea. BMC Vet Res. 2014;10:206. http://dx.doi.org/10.1186/s12917-014-0206-7 Address for correspondence: Caterina Guzmán-Verri, Programa de 13. Page AJ, Taylor B, Delaney AJ, Soares J, Seemann T, Keane JA, Investigación en Enfermedades Tropicales, Escuela de Medicina et al. SNP-sites: rapid efficient extraction of SNPs from multi-FASTA alignments. Microb Genomics. 2016 Apr 29. Veterinaria, Campus Benjamín Núñez, Universidad Nacional, Heredia http://dx.doi.org/10.1099/mgen.0.000056 Aptdo 304, 3000 Costa Rica; email [email protected]

@CDC_EIDjournal EID Follow the EID journal on Twitter and get the most current information journal from Emerging Infectious Diseases.

1000 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Isolated Case of Marburg Virus Disease, Kampala, Uganda, 2014

Luke Nyakarahuka, Joseph Ojwang, visited Python Cave in Queen Elizabeth National Park in Alex Tumusiime, Stephen Balinandi, 2008 (5,6). Both Python Cave and the Kitaka mine are in- Shannon Whitmer, Simon Kyazze, Sam Kasozi, habited by Egyptian fruit bats (Rousettus aegyptiacus), the Milton Wetaka, Issa Makumbi, Melissa Dahlke, host reservoir of MARV (7). Jeff Borchert, Julius Lutwama, Ute Ströher, In 2014, a fatal case of MVD occurred in Uganda. We Pierre E. Rollin, Stuart T. Nichol, report here on the field and laboratory investigation of this Trevor R. Shoemaker case, including possible sources of infection.

In September 2014, a single fatal case of Marburg virus The Study was identified in a healthcare worker in Kampala, Uganda. On September 23, 2014, a healthcare worker was admitted The source of infection was not identified, and no second- to Mengo Hospital in Kampala, Uganda, with a febrile ill- ary cases were identified. We describe the rapid identification, laboratory diagnosis, and case investigation of the third ness suspected to be viral hemorrhagic fever (VHF) infec- Marburg virus outbreak in Uganda. tion. The patient, a 30-year-old man, was a radiographer who worked at Mengo Hospital and part-time at Mpigi Health Center IV in the Mpigi district, ≈30 km south of Kampala arburg virus disease (MVD) is caused by Marburg vi- (Figure 1). His symptoms began on September 17, 2014. A Mrus (MARV; family Filoviridae, which also includes rapid diagnostic test result was positive for malaria, and the Ebola viruses). The disease was first discovered in 1967 in patient was given intravenous ceftriaxone, 5% dextrose, and Marburg and Frankfurt, Germany, after laboratory work- artesunate. However, his condition continued to deteriorate. ers were infected from monkeys imported from Uganda On September 26, he began to display hemorrhagic (1). Thereafter, sporadic cases and outbreaks of MVD have signs, notably profuse bleeding from body orifices. Clini- been documented in South Africa, Kenya, the Democratic cal findings included fever (temperature 38.4°C), nausea, Republic of the Congo, Angola, Uganda, the Netherlands, vomiting, diarrhea, musculoskeletal pain, abdominal pain, and the United States (2). MVD remains of great public headache, sore throat, difficulty swallowing, difficulty health importance because of the case-fatality rate, which breathing, anorexia, bleeding from the nose, bloody stools, can be as high as 90%, and documented human-to-human vomiting blood, and upper gastrointestinal tract bleeding. transmission, with associated socioeconomic consequences. He died on September 28. Uganda has experienced previous outbreaks of MVD, Whole blood and serum samples collected on Septem- resulting in fatalities and socioeconomic effects from loss ber 28 were sent to the Uganda Virus Research Institute of tourism. The first recorded outbreak in Uganda occurred (UVRI)/US Centers for Disease Control and Prevention in the Kamwenge district in 2007, where 4 MVD cases (CDC) VHF laboratory in Entebbe for testing on Septem- were confirmed in miners at the Kitaka mine (3). A second, ber 30. We performed diagnostic testing by real-time re- larger outbreak occurred in the western Uganda districts of verse transcription PCR (RT-PCR), antigen detection, and Kabale, Ibanda, and Kamwenge in 2012 (4) and was also MARV IgM ELISA on the suspected sample as described linked to mining activity in the Ibanda district. In addi- (8). We detected MARV by RT-PCR, but antigen detection tion, tourists from the United States and the Netherlands and IgM serologic results were negative. were infected with MARV in western Uganda when they We obtained independent confirmation from duplicate Author affiliations: Uganda Virus Research Institute, Entebbe, whole blood and serum samples by using RT-PCR primers Uganda (L. Nyakarahuka, J. Lutwama); US Centers for Disease and probes targeting MARV VP40 gene and a commercial Control and Prevention, Entebbe (J. Ojwang, A. Tumusiime, filovirus PCR screening assay (Altona Diagnostics, Ham- S. Balinandi, J. Borchert, T.R. Shoemaker); Centers for Disease burg, Germany) (9) (Table). On the basis of the positive Control and Prevention, Atlanta, Georgia, USA (S. Whitmer, RT-PCR results, we confirmed MVD in this patient. We U. Ströher, P.E. Rollin, S.T. Nichol); Public Health Emergency shipped specimens to the CDC Viral Special Pathogens Operations Center, Kampala, Uganda (S. Kyazze, S. Kasozi, Branch, Division of High-Consequence Pathogens and Pa- M. Wetaka, I. Makumbi, M. Dahlke) thology, National Center for Emerging and Zoonotic In- fectious Diseases (Atlanta, GA, USA), for further testing, DOI: https://dx.doi.org/10.3201/eid2306.170047 including virus isolation and sequencing. A virus isolate

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1001 DISPATCHES

Figure 1. Locations where patient with confirmed Marburg virus disease lived, worked, and was buried, Kampala, Uganda, 2014. Inset map shows location of Uganda in Africa.

(812601) was generated from the clinical specimen after a We created an outbreak database using the Epi Info single passage in cell culture (Vero E6). Viral Hemorrhagic Fever outbreak management appli- After laboratory confirmation, a multidisciplinary cation (10). All case and contact data were managed by team from UVRI and CDC-Uganda performed the ini- the Uganda Ministry of Health Public Health Emergency tial outbreak investigations at Mengo Hospital and Mpigi Operations Center. We identified 197 close contacts, who Health Center IV. The investigation team provided the out- were followed for 21 days. During the course of follow- break case definition, and details of case contacts were ob- up, 33 (16.2%) of the 197 contacts converted to suspected tained from both health facilities. An investigation was also case-patients by exhibiting clinically compatible signs or performed in Kasese District, where the health worker was symptoms matching the outbreak case definition. Blood taken for burial. In addition, an ecologic investigation was samples from suspected case-patients were tested at the conducted to identify roosting sites nearby for the potential UVRI/CDC laboratory in Entebbe; all were negative for presence of R. aegyptiacus bats. MARV by RT-PCR and serologic analysis.

1002 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Marburg Virus Disease, Kampala, Uganda

Table. Primers used to generate Marburg virus-specific cDNA fragments for whole-genome sequencing of isolates from Uganda Genome fragment Primer set Primer sequence, 5’→3’ Region amplified Region size, kbp 1 63F TGA CAT TGA GAC TTG TCA GTC 64-4998 4.9 MARB-4997R GCT TGA TTT CCT TCA CGC 2 3005F AAG TCA GCG AGG GGT TGA TGA CTG GAA AAG 2970- 6438 3.5 6426R TGC TAT GTT CCC TTC AGT GAA GAC 3 6101F AGA AAA CAG AAG ACG TCC ATC TGA TG 6065- 9405 3.3 9405R ACT TAA TGC TGC ACG AAG TGA TG 4 7567F TGG CCC TGG AAT IGA AGG ACT C 7548-10571 3.0 10571R AGC ATA TGA ACA ATA GAT C 5 36F GTA CCT CTA AGG AAA ATC ATG AAG 9979-15442 5.4 57R GTT GAT ATA ATT GCA CGT GTA GAT 6 12006F TTG CCA GAA GGA TAA AAG GAC AAA GAG 11953- 15486 3.5 15516R ATT TTG GAA GAT TAT ATT ACT ATC 7 15519F TGG ACG ATA GGA AAT CGA GCA C 15108- 19111 4.0 19155R TGG ACA CAC TAA AAA GAT G

Conclusions resulting in large outbreaks that are much harder to control, We describe the second single-case filovirus outbreak de- as was seen during the 2014–2015 Ebola virus outbreak in tected in Uganda; a case of infection with Sudan ebola- West Africa (12). This finding emphasizes the need for con- virus was reported in Luwero District in 2011 (11). The tinued surveillance and awareness even when other, more investigation was unable to identify a conclusive source of common pathogens are initially suspected. infection, including evidence of the natural reservoir host The finding of no secondary cases in this investigation near where the infected patient was working or residing or can be attributed, in part, to use of infection control prac- potential cases in persons who visited health centers before tices and personal protective equipment when first encoun- this case was confirmed. No secondary cases arose from tering a suspected VHF case. Because Uganda has experi- contact with the initial case-patient. enced 10 VHF outbreaks since 2011, increased awareness The patient tested positive for malaria but later tested and use of personal protective equipment and infection positive for MARV by RT-PCR and serologic analysis. control practices have greatly limited secondary transmis- These findings indicate that co-infection of viral hemorrhag- sion, especially in the healthcare setting. Routine use of ic fever (VHF) and other tropical infectious diseases can gloves, protective gowns, and chlorine is now more com- confound diagnosis and delay early detection, potentially mon in lower-level healthcare facilities in Uganda; these

Figure 2. Phylogenetic tree comparing complete or nearly complete Marburg virus (MARV) genomes sequenced from bat and human sources in Uganda. A consensus whole- genome sequence was assembled by mapping reads to the reference MARV sequence NC_001608 using CLC Genomics Workbench (Waltham, MA, USA). A phylogenetic tree was constructed using MEGA6.06 (http://www.megasoftware.net). Viral sequences acquired from human sources are in standard type, and viral sequences acquired from bats are italicized; the sequence from the human case-patient described in this study, MBG 812601 2014, is in bold. Evolutionary history was inferred using the maximum-likelihood method based on the Tamura-Nei model with MEGA6.06. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. Values at nodes represent bootstrap values following 1,000 replicates. Scale bar represents substitutions per site. GenBank accession numbers used in this tree are KP985768, JX458855.1, FJ750957.1, JX458852.1, JX458854.1, FJ750958.1, JX458856.1, JX458828.1, JX458826.1, JX458834.1, DQ447651.1, JX458846.1, AY358025.2, DQ447657.1, Z12132.1, NC_001608.3, Z29337.1, EF446132.1, JN408064.1, KC545388.1, KC545387.1, DQ447649.1, EF446131.1, DQ447652.1, FJ750956.1, FJ750955.1, and FJ750953.1.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1003

DISPATCHES protective products were used at both Mengo Hospital and investigation of viral hemorrhagic fevers, including Ebola and Mpigi Health Center IV. Marburg viruses, and he specializes in characterizing the natural The full genomic sequence of this MARV isolate ecology of hemorrhagic fever viruses. (812601; Genbank accession no. KP985768) falls into a cluster that consists mostly of MARV sequences isolat- References ed from bats. The closest related sequence was obtained 1. Siegert R, Shu HL, Slenczka W. Isolation and identification from MARV isolated from a juvenile male R. aegyptia- of the “Marburg virus” [in German]. Dtsch Med Wochenschr. cus bat (Q843) captured in August 2009 in Python Cave, 1968;93:604–12. http://dx.doi.org/10.1055/s-0028-1105103 Queen Elizabeth National Park (Figure 2). Other viral se- 2. Centers for Disease Control and Prevention. Chronology of Marburg hemorrhagic fever outbreaks; known cases and outbreaks quences from bat specimens in this clade were from bats of Marburg hemorrhagic fever, in chronological order [cited collected from either Python Cave or the Kitaka mine dur- 2014 Oct 29]. http://www.cdc.gov/vhf/marburg/resources/ ing 2007–2009. The most closely related human sequence outbreak-table.html (01Uga07) was from a miner who worked in the Kitaka 3. Adjemian J, Farnon EC, Tschioko F, Wamala JF, Byaruhanga E, Bwire GS, et al. Outbreak of Marburg hemorrhagic fever among mine in July 2007 (3,7). Because of the wide genetic di- miners in Kamwenge and Ibanda Districts, Uganda, 2007. J Infect versity of MARVs, there is no definitive way to identify Dis. 2011;204(Suppl 3):S796–9. http://dx.doi.org/10.1093/infdis/ where this patient may have become infected. jir312 We do not know why the patient did not transmit 4. Knust B, Schafer IJ, Wamala J, Nyakarahuka L, Okot C, Shoemaker T, et al. Multidistrict outbreak of Marburg virus MARV to any of his close contacts. We can assume that disease—Uganda, 2012. J Infect Dis. 2015;212(Suppl 2):S119–28. none of the contacts had substantial exposure to the patient http://dx.doi.org/10.1093/infdis/jiv351 while he was infectious. The relatively small size of this 5. Centers for Disease Control and Prevention (CDC). Imported case and previous filovirus outbreaks in Uganda can be attribut- of Marburg hemorrhagic fever—Colorado, 2008. MMWR Morb Mortal Wkly Rep. 2009;58:1377–81. ed to enhanced VHF surveillance, rapid case identification, 6. Timen A, Koopmans MP, Vossen AC, van Doornum GJ, laboratory testing, and investments from the global health Günther S, van den Berkmortel F, et al. Response to imported case security agenda in rapid sample transportation to the na- of Marburg hemorrhagic fever, the Netherlands. Emerg Infect Dis. tional VHF reference laboratory for diagnostic testing (13). 2009;15:1171–5. http://dx.doi.org/10.3201/eid1508.090015 7. Towner JS, Amman BR, Sealy TK, Carroll SA, Comer JA, VHF surveillance continues to be a top priority for Uganda, Kemp A, et al. Isolation of genetically diverse Marburg viruses and the VHF surveillance program continues to play a cru- from Egyptian fruit bats. PLoS Pathog. 2009;5:e1000536. cial role in detecting, responding to, and helping to control http://dx.doi.org/10.1371/journal.ppat.1000536 these outbreaks. 8. Ksiazek TG, West CP, Rollin PE, Jahrling PB, Peters CJ. ELISA for the detection of antibodies to Ebola viruses. J Infect Dis. 1999;179(Suppl 1):S192–8. http://dx.doi.org/10.1086/514313 Acknowledgments 9. Diagnostics A. RealStar Filovirus Screen RT-PCR Kit 1.0, 2014 We thank the staff at Mengo Hospital, Kampala, and Mpigi [cited 2016 Jun 9]. http://www.altona-diagnostics.com/tl_files/ Health Center IV; the initial investigation team, which included website/downloads/RealStar_Filovirus Screen RT-PCR Kit 10_CE_ WEB_2014-08-29.pdf members from UVRI and CDC-Uganda for field and laboratory 10. Schafer IJ, Knudsen E, McNamara LA, Agnihotri S, Rollin PE, investigation of the outbreak; the Ministry of Health National Islam A. The Epi Info Viral Hemorrhagic Fever (VHF) Task Force on Epidemic Preparedness and Response; World application: a resource for outbreak data management and contact Health Organization; Kampala City Council Authority; Kasese tracing in the 2014–2016 West Africa Ebola epidemic. J Infect Dis. 2016;214(suppl3):S122–36. http://dx.doi.org/10.1093.infdis/jiw272 and Mpigi local governments; CDC Uganda, CDC, and UVRI 11. Shoemaker T, MacNeil A, Balinandi S, Campbell S, Wamala JF, Motorpool for transportation; and Médecins Sans Frontières for McMullan LK, et al. Reemerging Sudan Ebola virus disease establishing an isolation ward at Mulago Hospital. We especially in Uganda, 2011. Emerg Infect Dis. 2012;18:1480–3. thank UVRI for its support of the VHF surveillance program and http://dx.doi.org/10.3201/eid1809.111536 12. The Lancet. 1 year on—lessons from the Ebola outbreak for diagnostics laboratory. We also thank the United States Agency WHO. Lancet. 2015;385:1152. http://dx.doi.org/10.1016/ for International Development, African Field Epidemiology S0140-6736(15)60619-5 Network, and the United Nations Children’s Fund, which all 13. Borchert JN, Tappero JW, Downing R, Shoemaker T, participated and contributed to the Marburg outbreak response Behumbiize P, Aceng J, et al.; Centers for Disease Control and Prevention (CDC). Rapidly building global health security and provided support to the Uganda Ministry of Health. capacity—Uganda demonstration project, 2013. MMWR Morb Mortal Wkly Rep. 2014;63:73–6. Dr. Nyakarahuka is a zoonotic disease epidemiologist at the Uganda Virus Research Institute in Entebbe, Uganda, with a Address for correspondence: Trevor R. Shoemaker, Centers for Disease background in veterinary medicine and public health. His work Control and Prevention, P.O. Box 7007, Plot 1577 Ggaba Road, specifically focuses on the surveillance, detection and Kampala, Uganda; email: [email protected]

1004 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Crimean-Congo Hemorrhagic Fever in Migrant Worker Returning from Oman to India, 2016

Pragya D. Yadav, Sachin Thacker, Deepak Y. Patil, aminotransaminase levels of 130 U/L (reference range 0–35 Rajlaxmi Jain, Devendra T. Mourya U/L) and 240 U/L (reference range 3–36 U/L), respectively; and prothrombin time of 17.9 sec (reference range 10–13 In January 2016, a migrant worker who returned home to sec). Results for brain multidetector computed tomography India after becoming ill in Oman was confirmed to have scanning were normal. The patient was administered intra- Crimean-Congo hemorrhagic fever (CCHF). Physicians venous ceftriaxone, pantoprazole, somazina citicoline, and should include CCHF in the differential diagnosis for patients with hemorrhagic signs and a history of recent travel cerebroprotein hydrolysate and oral ribavirin. to any area where CCHF is endemic or prevalent. On illness day 5, the physician sent a clinical sample to the National Institute of Virology in Pune, India, for testing. Results were negative for dengue NS1 antigen and ncreased international travel has led to the global spread IgM. However, using real-time reverse transcription PCR Iof numerous diseases (1). One of these diseases, Crime- and IgM ELISA as previously described (4), the Institute an-Congo hemorrhagic fever (CCHF), has affected persons detected CCHF viral RNA (9.0 × 102 copies/5 μL) and IgM. in >20 countries in Africa, Asia, southeastern Europe, and After confirming CCHFV infection in the patient, the the Middle East (2). Transmission of CCHF virus (CCH- hospital placed him in quarantine and implemented strict FV) to humans primarily occurs via Hyalomma spp. ticks barrier nursing practices; no secondary cases occurred. and livestock. Large numbers of nosocomial and sporadic Contacts in India were closely monitored for 15 days; all CCHF outbreaks have been reported in humans worldwide, remained asymptomatic. The patient was discharged 13 including in India, where information about travel-asso- days after illness onset. A clinical sample tested 15 days ciated CCHF cases is lacking (3). We report on a case of after discharge was positive for CCHFV IgG and IgM and CCHF in a man who returned home to India after becoming negative for viral RNA. ill in Oman. We attempted to isolate CCHFV by inoculating infant CD-1 mice with 20 μL of serum or blood collected from the The Study patient on illness day 5. The mice showed no clinical signs On January 24, 2016, a 33-year-old migrant worker from by postinoculation day 7, when we euthanized the mice, India became ill while working in Muscat, Oman. He ex- harvested brains, made a brain suspension in 1.25% bovine perienced abdominal pain, occasional dysuria, anorexia, serum albumin in phosphate-buffered saline, and used it to nausea, and vomiting. The man, a supervisor on an animal inoculate infant mice. Beginning on postinoculation day farm, had occasional contact with animals, including cows, 5, these mice began showing neurologic signs, (e.g., hind goats, horses, and camels. On January 26, he was admitted limb paralysis, circling movement); 1 mouse died. We also to a hospital in Muscat and diagnosed with severe thrombo- attempted to isolate CCHFV by inoculating the brain sup- cytopenia and acalculous cholecystitis; he was discharged pension into Vero CCL-81 cells. We detected high virus with a referral to the government hospital for further care. loads in the mouse brain suspensions (1.2 × 104 copies/5 On January 27, instead of visiting the government hospi- μL) and tissue culture fluid (2.3× 104 copies/5 μL) (4). tal, he traveled to his residence in Gujarat State, India. On Using the cell culture–grown virus and previously January 28, he was admitted to a hospital in Kutch District, established protocols (5), we sequenced the complete ge- Gujarat, with fever; hemorrhagic signs (melena, epistaxis, nome (small [S], large [L], and medium [M] segments) and hematuria); vomiting; loss of appetite; and altered sen- of the virus. We used Sequencher 5.4.5 DNA Sequence 9 sorium. He had a platelet count of 33 × 10 /L (reference Analysis Software (Gene Codes Corporation, Ann Arbor, 9 range 130–400 × 10 /L); hemoglobin level of 6.8 g/dL MI, USA) to align sequences and submitted them to Gen- (reference range 14.0–17.4 g/dL); aspartate and alanine Bank (accession nos. KY213714 [S gene], KY213712 [L gene], KY213713 [M gene]). We used a neighbor-joining Authors affiliations: National Institute of Virology, Pune, India algorithm in MEGA6 (http://www.megasoftware.net/) to (P.D. Yadav, D.Y. Patil, R. Jain, D.T. Mourya); JK Hospital, Bhuj, perform phylogenetic analyses, and we constructed a phy- India (S. Thacker) logenetic tree using the sequences of the 3 segments and previously identified representative CCHFV strains. DOI: https://dx.doi.org/10.3201/eid2306.161950

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1005 DISPATCHES

Figure 1. Phylogenetic tree comparing the small gene segment of Crimean- Congo hemorrhagic fever virus (CCHFV) strain isolated in India (bold) with reference CCHFV strains obtained from GenBank. The strain from India, NIV161064, was isolated in 2016 from the serum of a patient who had returned home to India after becoming ill in Oman. Representative reference strains were inferred by the neighbor-joining algorithm in MEGA6 (http://www.megasoftware. net/). Strains are identified by GenBank accession number and location and year of isolation. CCHFV groups are shown on the right. Scale bar indicates number of nucleotide substitutions per site.

1006 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 CCHF in Migrant Worker Returning to India

By analyzing the S segment, we demonstrated that cross-sectional serosurvey that showed CCHFV is prev- this CCHFV strain belongs to the Middle East Asia group alent among livestock in 22 states and a union territory IV (Asia-1) of CCHFVs, along with strains from Iran, Af- of India (9). ghanistan, Pakistan, and Oman (Figure 1); the S segment Our findings show that the S segment of the CCHFV was closest to those of strains isolated in 2007 in Zehedan, isolated in this study shared maximum relatedness with Iran (GenBank accession no. KC867274). The L segment the Middle East Asia group IV isolates, and the M seg- also belongs to the Asia-IV group, along with strains from ment belongs to the M2 group, which also includes strains Afghanistan, Tajikistan, Iraq, and Oman and strains iso- from countries in the Middle East. Thus, the strain isolated in India in 2011 (Figure 2, panel A); the L segment lated from the migrant worker is a combination of the S was closest to that of a strain isolated in 2012 in Afghani- gene of Asia-I (S–Asia-1) and M2 strains. We also ob- stan (GenBank accession no. KC344855). The M segment served parallel clustering of the S and L segments with belongs to type M2 and was closest to the M segment of Asia-IV group viruses. Like the strain in this study, most strains isolated during 2004–2007 in Iran (GenBank acces- CCHFVs circulating in the Middle East are a combination sion nos. DQ446216, DQ446215, and KC867273) (Figure of S–Asia-1 and M2 strains (12,13). However, CCHFV 2, panel B). strains reported from different districts of Gujarat- dur ing 2011–2015 were a combination of S–Asia-2 and Far Conclusion East M2 viruses that had different parental origins in the During 2011–2016, many CCHF cases were reported S (from Tajikistan strain TADJ/HU8966) and L and M from India, especially from Gujarat State, where the vi- (from Afghanistan strain Afg09-2990) segments, suggest- rus has been endemic since 2011, and adjoining Rajsthan ing that an intragenotypic reassortant sequence entered State, where a few sporadic cases have occurred (4,6–11). into India. CCHF cases have been reported from Oman Furthermore, we previously conducted a countrywide, since 1995 (14,15). However, because sequences have not

Figure 2. Phylogenetic trees comparing the large (A) and medium (B) gene segments of Crimean-Congo hemorrhagic fever virus (CCHFV) strain isolated in India (bold) with reference CCHFV strains obtained from GenBank. The strain from India, NIV161064, was isolated in 2016 from the serum of a patient who had returned home to India after becoming ill in Oman. Representative reference strains were inferred by the neighbor-joining algorithm in MEGA6 (http://www.megasoftware.net/). Strains are identified by GenBank accession number and location and year of isolation. CCHFV groups are shown on the right. Scale bars indicate number of nucleotide substitutions per site.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1007 DISPATCHES been reported for recent strains, we could not conduct a 3. Leblebicioglu H, Ozaras R, Fletcher TE, Beeching NJ; more robust phylogenetic analysis. Overall, our phyloge- ESCMID Study Group for Infections in Travellers and Migrants (ESGITM). Crimean-Congo haemorrhagic fever in travellers: a netic analyses substantiate that the case-patient in Gujarat systematic review. Travel Med Infect Dis. 2016;14:73–80. was infected with a CCHFV strain from the Middle East http://dx.doi.org/10.1016/j.tmaid.2016.03.002 while working in Oman. 4. Mourya DT, Yadav PD, Shete AM, Gurav YK, Raut CG, Jadi RS, The case-patient we report became ill while in Oman et al. Detection, isolation and confirmation of Crimean-Congo hemorrhagic fever virus in human, ticks and animals in and traveled to Gujarat within the incubation period for Ahmadabad, India, 2010–2011. PLoS Negl Trop Dis. CCHFV (6–7 days). On illness day 5, the case-patient 2012;6:e1653. http://dx.doi.org/10.1371/journal.pntd.0001653 was hospitalized in Gujarat and confirmed to be infected 5. Yadav PD, Cherian SS, Zawar D, Kokate P, Gunjikar R, Jadhav S, with CCHFV. Results of serologic testing for IgM cor- et al. Genetic characterization and molecular clock analyses of the Crimean-Congo hemorrhagic fever virus from human and ticks in roborate that the patient acquired the infection while India, 2010–2011. Infect Genet Evol. 2013;14:223–31. in Oman. http://dx.doi.org/10.1016/j.meegid.2012.10.005 Many reports have been made around the world of 6. Yadav PD, Raut CG, Mourya DT. Re-occurrence of Crimean- travelers inadvertently importing diseases from one coun- Congo haemorrhagic fever in Ahmedabad, Gujarat, India (2012): a fatal case report. Indian J Med Res. 2013;138:1027–8. try to another. Thus, travelers should be made aware of 7. Yadav PD, Gurav YK, Mistry M, Shete AM, Sarkale P, communicable diseases present in countries they visit, and Deoshatwar AR, et al. Emergence of Crimean-Congo hemorrhagic patients should inform doctors if they have a recent travel fever in Amreli District of Gujarat State, India, June to July 2013. history. In addition, physicians should consider CCHF in Int J Infect Dis. 2014;18:97–100. http://dx.doi.org/10.1016/ j.ijid.2013.09.019 the differential diagnosis of patients who have hemorrhagic 8. Mourya DT, Yadav PD, Shete A, Majumdar TD, Kanani A, signs and have recently returned from any area where CCH- Kapadia D, et al. Serosurvey of Crimean-Congo hemorrhagic fever FV is endemic or prevalent. Increased international travel virus in domestic animals, Gujarat, India, 2013. Vector Borne will result in further importations of infectious diseases, Zoonotic Dis. 2014;14:690–2. http://dx.doi.org/10.1089/ vbz.2014.1586 highlighting the need for worldwide disease surveillance 9. Mourya DT, Yadav PD, Shete AM, Sathe PS, Sarkale PC, and for implementation of the World Health Organization Pattnaik B, et al. Cross-sectional serosurvey of Crimean-Congo International Health Regulations (http://www.who.int/ hemorrhagic fever virus IgG in livestock, India, 2013–2014. topics/international_health_regulations/en/). Emerg Infect Dis. 2015;21:1837–9. http://dx.doi.org/10.3201/ eid2110.141961 10. Makwana D, Yadav PD, Kelaiya A, Mourya DT. First confirmed Acknowledgment case of Crimean-Congo haemorrhagic fever from Sirohi district in We thank JK Hospital, Bhuj, Gujarat, India, for sharing clinical Rajasthan State, India. Indian J Med Res. 2015;142:489–91. samples. We also thank Anita Shete, Vimal Kumar, Rajen Lakra, http://dx.doi.org/10.4103/0971-5916.169221 11. Yadav PD, Patil DY, Shete AM, Kokate P, Goyal P, Jadhav S, Prasad Sarkale, Divya Bhattad, Pravin Kore, Kumar Bagmare, et al. Nosocomial infection of CCHF among health care and Shital Dalal for technical support; and we are sincerely workers in Rajasthan, India. BMC Infect Dis. 2016;16:624. grateful to Sarah Cherian for sharing her expertise on http://dx.doi.org/10.1186/s12879-016-1971-7 genome analysis. 12. Deyde VM, Khristova ML, Rollin PE, Ksiazek TG, Nichol ST. Crimean-Congo hemorrhagic fever virus genomics and global Financial support was provided by the National Institute of diversity. J Virol. 2006;80:8834–42. http://dx.doi.org/10.1128/ JVI.00752-06 Virology, Pune, India. 13. Lukashev AN. Evidence for recombination in Crimean-Congo Dr. Yadav is a group leader for the maximum containment hemorrhagic fever virus. J Gen Virol. 2005;86:2333–8. http://dx.doi.org/10.1099/vir.0.80974-0 laboratory, National Institute of Virology, Pune, India. Her 14. Schwarz TF, Nitschko H, Jäger G, Nsanze H, Longson M, Pugh RN, research interest is zoonotic diseases of high risk group viruses et al. Crimean-Congo haemorrhagic fever in Oman. Lancet. including CCHF, Kyasanur Forest disease, and Nipah virus. 1995;346:1230. http://dx.doi.org/10.1016/S0140-6736(95)92936-3 15. Body MHH, Abdulmajeed HA, Hammad MH, Mohamed SA, Saif SA, AL-Maklady S, et al. Cross-sectional survey of References Crimean-Congo hemorrhagic fever virus in the sultanate of Oman. 1. Tatem AJ, Rogers DJ, Hay SI. Global transport networks and Journal of Veterinary Medicine and Animal Health. 2016;8:44–9. infectious disease spread. Adv Parasitol. 2006;62:293–343. http://dx.doi.org/10.5897/JVMAH2016.0472 http://dx.doi.org/10.1016/S0065-308X(05)62009-X 2. Ince Y, Yasa C, Metin M, Sonmez M, Meram E, Benkli B, et al. Address for correspondence: Devendra T. Mourya, National Institute Crimean-Congo hemorrhagic fever infections reported by ProMED. Int J Infect Dis. 2014;26:44–6. http://dx.doi.org/10.1016/j. of Virology, 20-A, Dr. Ambedkar Rd, Pune, 411 001, India; email: ijid.2014.04.005 [email protected] or [email protected]

1008 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Rise in Group W Meningococcal Carriage in University Students, United Kingdom

Neil J. Oldfield, Caroline Cayrou, university students (8). A catch-up campaign was used to Mahab A.K. AlJannat, Ali A.A. Al-Rubaiawi, offer MenACWY vaccine to all 14–18-year-old persons, Luke R. Green, Shehzan Dada, Oliver D. Steels, with persons who left school in 2015 (17–18 years of age) Christopher Stirrup, Joe Wanford, given priority. Banan A.Y. Atwah, Christopher D. Bayliss, The rationale for targeting the vaccine to older ad- David P.J. Turner olescents and young adults stems from carriage studies showing that this demographic represents the principal MenACWY conjugate vaccination was recently introduced reservoir of meningococcal carriage, as well as experi- in the United Kingdom for adolescents and young adults ence with other polysaccharide conjugate vaccines. In- to reduce disease from infection by Neisseria meningitidis troduction of the MenC monovalent conjugate vaccines group W. We conducted a cross-sectional meningococcal carriage study in first-year UK university students. Despite previously reduced carriage acquisition of MenC strains 71% MenACWY vaccine coverage, carriage of group W in- in adolescents and young adults (9). There is evidence, creased substantially. albeit limited, to suggest that the quadrivalent vaccine may have a similar effect on carriage of MenCWY strains (10,11). Reduced carriage in this population should lead eisseria meningitidis causes severe sepsis and men- to indirect protection of other age groups through herd Ningitis. The main reservoir in most populations is as- immunity, thus enhancing the public health effect and ymptomatic pharyngeal carriage in older adolescents and cost-effectiveness of this approach. young adults (1). High carriage rates are particularly evident in semiclosed communities of young adults, such as The Study university student populations, where persons live, work, To assess trends in meningococcal strain carriage, and to and socialize together (2). determine the immediate effect of the MenACWY vac- Meningococcal carriage was previously assessed in cine on carriage of MenW/Y strains, we conducted a university students in the United Kingdom during 2009–10 cross-sectional study in first-year students at the UoN from at the University of Nottingham (UoN) when a high preva- registration during September 2015 through March 2016. lence of capsular group Y (MenY) meningococcal carriage MenACWY vaccination coverage in this student popula- was detected (3). This high level of MenY carriage was tion increased from 31% (preregistration) to ≈70% (im- concomitant with a rise in disease caused by MenY strains mediately postregistration) as a result of a campus-based in the United Kingdom (4). Since 2009, although MenY vaccination campaign targeting freshmen (12). disease has plateaued, capsular group W (MenW) disease The study was approved by the Research Ethics Com- has steadily increased because of the rapid expansion of hy- mittee at the UoN, and written informed consent was ob- pervirulent strains belonging to the sequence type 11 clonal tained from all participants. We recruited convenience complex (MenW:ST-11) (5). samples of first-year students in September and November Analysis of whole-genome sequence data has shown 2015 and March 2016. In September, students were recruit- that isolates from the same MenW:ST-11 lineage, termed ed during registration; in November and March, students the South American/UK strain, are also endemic to Chile, were recruited in 5 dormitories (A–E) with single-occupan- Brazil, and Argentina (6) and were recently reported in cy rooms. We searched the UoN Health Service registra- Australia (7). In response, Public Health England intro- tion database (EMIS Web software; EMIS Health, Leeds, duced the MenACWY vaccine in the routine adolescent UK) to determine vaccination status in registered first-year school program for 14- and 15-year-olds and first-year students on arrival at the UoN and after the campus-based vaccination campaign. Author affiliations: University of Nottingham, Nottingham, UK In September, we obtained pharyngeal swab speci- (N.J. Oldfield, M.A.K. AlJannat, O.D. Steels, C. Stirrup, B.A.Y. mens from eligible students immediately before they Atwah, D.P.J. Turner); University of Leicester, Leicester, UK received the MenACWY vaccine. We immediately in- (C. Cayrou, A.A.A. Al-Rubaiawi, L.R. Green, S. Dada, J. Wanford, oculated all pharyngeal swabs onto GC selective agar C.D. Bayliss) (Oxoid, Basingstoke, UK) and incubated them at 37°C DOI: http://dx.doi.org/10.3201/eid2306.161768 in air containing 5% CO2. After 24 and 48 hours, we

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1009 DISPATCHES selected oxidase-positive colonies suggestive of Neisseria United Kingdom (5). We examined capsular expression spp. and confirmed their identity by amplification of the by serogrouping and found 32 (62%) of the MenW iso- meningococcal gene crgA plus ctrA and/or porA, as de- lates expressed the W capsular polysaccharide. Of the 21 scribed previously (13). MenW carriers in March 2016 who had been vaccinated We performed PCR-based genogrouping as described at least 5 months before sampling, 15 (71%) were harbor- previously (13,14). The Meningococcal Reference Unit, ing isolates expressing the W capsule (online Technical Public Health England, Manchester, UK, performed sero- Appendix Table 2). grouping and serotyping of MenW isolates using dot-blot enzyme immunoassay. Sequence data derived from ampli- Conclusions fied porA and factor H–binding protein alleles was queried We detected a rapid rise in carriage of MenW among first- against the PubMLST/Neisseria database (http://pubmlst. year students in a university setting in the United King- org/neisseria). We performed χ2 tests for significance by dom. In comparison, carriage of MenC in adolescents and using STATCALC (Epi Info version 7.2.0.1; Centers for young adults in the United Kingdom, including in univer- Disease Control and Prevention, Atlanta, GA, USA). sity students, was rare before the introduction of MenC The September sample of 769 first-year students -rep monovalent conjugate vaccines (9). The rise in MenW resented 10.9% of the 7,049 first-year students registered in carriage is most likely due to acquisition within student 2015. Carriage rates among this group increased from 14% dormitories or social spaces on campus but was unexpect- in late September 2015 to 39% by mid-November 2015 and ed because no such isolates were found in a similar study 46% in March 2016 (online Technical Appendix Table 1, at the UoN in 2008–09 (15), and only a limited number in http://wwwnc.cdc.gov/EID/article/23/6/16-1768-Techapp1. a multicenter carriage study involving UK universities in pdf). The characteristics of enrolled students and behavioral 2010–11 (6,10). risk factors for carriage were similar at the 3 time points. The The increase in MenW carriage among university stu- initial carriage rate of 14% was much lower for first-year dents merits further monitoring because it could contrib- students in September 2015 than in September 2009 (23.2%; ute further to the sustained increase in MenW disease in χ2 = 34, df = 1; p<0.00001) (3), suggesting a reduction in me- the United Kingdom. Two cases of MenW disease were ningococcal carriage in adolescents in the United Kingdom, reported in unvaccinated students in Nottingham during possibly attributable to an alteration in risk factors for car- 2015–16. Students attending universities exhibit high mo- riage. The MenY carriage rate for incoming students (1.8%) bility and may represent an ongoing vehicle for amplifica- was also lower than that detected in 2009 (2.9%; χ2 = 2.0, df tion and spread of MenW into communities throughout the = 1; p = 0.15) (3). United Kingdom or beyond, with implications for vaccina- In September 2015, carriage rates were 4.2% for geno- tion policy and future research. group capsule null locus (cnl), 3.3% for B, 1.8% for Y, and 0.7% for W strains. A substantial part of the increase Acknowledgments in carriage from September 2015 through March 2016 was We are grateful to all participants in this study. We thank the result of a notable increase (0.7% to 8.0%) in carriage Daniel Hammersley for logistical assistance and for providing of MenW strains (online Technical Appendix Table 1). MenACWY vaccination data. We also thank Steve Gray, No change in the carriage of MenY strains was detected Anthony Carr, and Ray Borrow for serogrouping and serotyping (online Technical Appendix Table 1). Of the 50 students meningococcal isolates. colonized with MenW, 36 (72%) had received MenACWY This work was supported by the University of Nottingham. vaccine either before or during registration, which is consistent with the overall MenACWY coverage in our first- Dr. Oldfield is a senior postdoctoral researcher at the School of year student cohort. Students colonized with MenW at the Life Sciences, University of Nottingham, Nottingham, UK. His second and third sampling times were distributed across research interests include meningococcal carriage dynamics all 5 dormitories, suggesting widespread dissemination, and pathogenesis. and 21 (91%) of the MenW carriers in the last time point (March 2016) had been vaccinated at least 5 months before References sampling (online Technical Appendix Table 2). 1. Caugant DA, Maiden MCJ. Meningococcal carriage and disease— Analysis of the genogroup W isolates showed that 47 population biology and evolution. Vaccine. 2009; 27(Suppl 2):B64– (90%) of 52 were serotype 2a (online Technical Appendix 70. http://dx.doi.org/10.1016/j.vaccine. 2009.04.061 Table 1) and harbored alleles for factor H binding protein 2. Neal KR, Nguyen-Van-Tam JS, Jeffrey N, Slack RC, Madeley RJ, Ait-Tahar K, et al. Changing carriage rate of Neisseria meningitidis peptide 22 and PorA P1.5,2, identical to the correspond- among university students during the first week of term: cross ing alleles harbored by the MenW:ST-11 clone respon- sectional study. BMJ. 2000;320:846–9. http://dx.doi.org/10.1136/ sible for the increase in invasive MenW disease in the bmj.320.7238.846

1010 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Group W Meningococcal Carriage in Students, UK

3. Ala’Aldeen DAA, Oldfield NJ, Bidmos A,F Abouseada NM, Ahmed NW, Turner DPJ, et al. Carriage of meningococci by university students, United Kingdom. Emerg Infect Dis. EID Podcast: 2011;17:1762–3. http://dx.doi.org/10.3201/eid1709.101762 4. Ladhani SN, Lucidarme J, Newbold LS, Gray SJ, Carr AD, Stained Glass and Flu Findlow J, et al. Invasive meningococcal capsular group Y disease, England and Wales, 2007–2009. Emerg Infect Dis. 2012;18:63–70. http://dx.doi.org/10.3201/eid1801.110901 5. Ladhani SN, Beebeejaun K, Lucidarme J, Campbell H, Gray S, Kaczmarski E, et al. Increase in endemic Neisseria meningitidis capsular group W sequence type 11 complex associated with severe invasive disease in England and Wales. Clin Infect Dis. 2015;60:578–85. http://dx.doi.org/10.1093/cid/ciu881 6. Lucidarme J, Hill DMC, Bratcher HB, Gray SJ, du Plessis M, Tsang RS, et al. Genomic resolution of an aggressive, widespread, diverse and expanding meningococcal serogroup B, C and W lineage. J Infect. 2015;71:544–52. http://dx.doi.org/10.1016/ j.jinf.2015.07.007 7. Carville KS, Stevens K, Sohail A, Franklin LJ, Bond KA, Brahmi A, et al. Increase in meningococcal serogroup W disease, Victoria, Australia, 2013–2015. Emerg Infect Dis. 2016;22:1785–7. http://dx.doi.org/10.3201/eid2210.151935 8. Campbell H, Saliba V, Borrow R, Ramsay M, Ladhani SN. Targeted vaccination of teenagers following continued rapid endemic expansion of a single meningococcal group W clone (sequence type 11 clonal complex), United Kingdom 2015. Euro Surveill. 2015;20:21188. http://dx.doi.org/10.2807/1560-7917. ES2015.20.28.21188 9. Maiden MCJ, Ibarz-Pavón AB, Urwin R, Gray SJ, Andrews NJ, Clarke SC, et al. Impact of meningococcal serogroup C conjugate vaccines on carriage and herd immunity. J Infect Dis. 2008; 197:737–43. http://dx.doi.org/10.1086/527401 10. Read RC, Baxter D, Chadwick DR, Faust SN, Finn A, Gordon SB, et al. Effect of a quadrivalent meningococcal ACWY glycoconjugate or a serogroup B meningococcal vaccine on meningococcal carriage: an observer-blind, phase 3 randomised clinical trial. Lancet. 2014;384:2123–31. http://dx.doi.org/10.1016/ S0140-6736(14)60842-4 11. Korzeniewski K, Skoczyńska A, Guzek A, Konior M, Chciałowski A, Waśko I, et al. Effectiveness of immunoprophylaxis in suppressing carriage of Neisseria meningitidis in the military environment. Adv Exp Med Biol. 2014;836:19–28. http://dx.doi.org/10.1007/5584_2014_22 12. Turner DPJ, Oldfield NJ, Bayliss CD. University vaccine campaign increases meningococcal ACWY vaccine coverage. Public Health. 2017;145:1–3. http://dx.doi.org/10.1016/j. puhe.2016.12.010 The work of art shown here depicts the interrela- 13. Taha M-K, Alonso J-M, Cafferkey M, Caugant DA, Clarke SC, tionship of human, animal, and environmental health. Diggle MA, et al. Interlaboratory comparison of PCR-based identification and genogrouping ofNeisseria meningitidis. Stained-glass windows have been appreciated J Clin Microbiol. 2005;43:144–9. http://dx.doi.org/10.1128/ for their utility and splendor for more than 1,000 JCM.43.1.144-149.2005 years, and this engaging work of art by stained glass 14. Claus H, Maiden MCJ, Maag R, Frosch M, Vogel U. Many carried meningococci lack the genes required for capsule synthesis artist Jenny Hammond reminds us that influenza and transport. Microbiology. 2002;148:1813–9. http://dx.doi.org/ A viruses—which can be easily spread between 10.1099/00221287-148-6-1813 animals and human, use various host species, and 15. Bidmos FA, Neal KR, Oldfield NJ, Turner DPJ, Ala’Aldeen DAA, Bayliss CD. . Persistence, replacement, and rapid clonal expansion exist in many different environments—remain an of meningococcal carriage isolates in a 2008 university student enduring and global health concern. cohort. J Clin Microbiol. 2011;49:506–12. http://dx.doi.org/ 10.1128/JCM.01322-10 Visit our website to listen: Address for correspondence: David P.J. Turner, Centre for Biomolecular https://www2c.cdc.gov/ Sciences, School of Life Sciences, University of Nottingham, podcasts/player. Nottingham NG7 2RD, UK; email: [email protected] asp?f=8644950

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1011 DISPATCHES

Penicillin Resistance of Nonvaccine Type Pneumococcus before and after PCV13 Introduction, United States

Cheryl P. Andam,1 Colin J. Worby,1 differences in antibiotic use, leading to differences in the -in Ryan Gierke, Lesley McGee, Tamara Pilishvili, tensity of selective pressure acting on the bacterial popula- William P. Hanage tion (9). The variation in the proportion of resistant isolates within individual serotypes in the United States was thought Introduction of 13-valent pneumococcal conjugate vaccine to be a reflection of this regional difference in antibiotic use in the United States was not associated with a significant and was identified as the major factor in driving geographic change in prevalence of penicillin resistance in nonvaccine variation of penicillin resistance (7). However, post-PCV7, type serotypes because of the variable success of highly resistant serotypes. Differences in regional serotype distri- this factor played a diminishing role in explaining geo- bution and serotype-specific resistance contributed to geographic heterogeneity in penicillin resistance, with variation graphic heterogeneity of penicillin resistance. in serotype distribution between sites being of increasing importance (8). Understanding the underlying causes of the geographic heterogeneity of penicillin resistance and the role treptococcus pneumoniae (pneumococcus) causes a of selective pressure provides important insights on the long- Srange of debilitating and potentially life-threatening in- term dynamics of penicillin resistance in the United States. fections, such as pneumonia, meningitis, and septicemia. To reduce illness and death caused by pneumococcal diseases, a The Study 7-valent pneumococcal conjugate vaccine (PCV7) was intro- To analyze NVT penicillin-nonsusceptible pneumococcus duced in 2000 and targeted serotypes 4, 6B, 9V, 14, 18C, 19F, (PNSP) detected in patients with invasive pneumococcal dis- and 23F. However, although vaccine type serotypes declined ease, we used data from the Active Bacterial Core surveillance in frequency after PCV7 introduction (1,2), an increasing fre- (ABCs) system, a population- and laboratory-based collab- quency of nonvaccine type (NVT) serotypes in samples from orative system between the Centers for Disease Control and carriage and invasive disease was observed in subsequent Prevention and state health departments and academic institu- years (2,3). Known as serotype replacement, this population- tions in 10 states (California, Colorado, Connecticut, Georgia, level change in serotype distribution, which most often in- Maryland, Minnesota, New Mexico, New York, Oregon, and volves preexisting clones and serotypes that were already in Tennessee). We considered PNSP non-PCV13 serotypes de- circulation before vaccine implementation (4), can reduce tected in patients in all age groups from 2009 (pre-PCV13, n = the benefits of vaccination 5( ). To address the rise in inva- 285 patients) through 2012 (post-PCV13, 339 patients). Non- sive pneumococcal disease associated with NVT serotypes, susceptibility was based on the meningitis breakpoint (MIC a second-generation conjugate vaccine was implemented in >0.12 µg/mL), as recommended by the Clinical and Labora- 2010 (PCV13), targeting the 7 serotypes targeted by PCV7 tory Standards Institute (10). Serotypes 15B, 15C, and 15B/C plus 6 additional serotypes: 1, 3, 5, 6A, 7F, and 19A (6). were grouped together as 15BC because of the reported re- The prevalence of penicillin-resistant pneumococcus versible switching between the 2 serotypes, which makes the strains varies considerably between states (7,8). Such varia- precise differentiation of these serotypes difficult11 ( ). tion might be caused by differences in serotype distribution To determine whether geographic differences in the (such that some locations have a higher prevalence of strains proportions of PNSP were consistent across serotypes, we that are generally more resistant) or higher-than-average calculated the proportions of PNSP for each of the 7 most levels of resistance within serotypes. Before the introduc- common NVT serotypes (15A, 15BC, 16F, 23A, 23B, tion of PCV7, regional variations in the prevalence of an- 33F, and 35B) across the 10 sites for 2009 and 2012. We tibiotic resistance were considered to be caused by regional found that serotypes with the highest proportions of PNSP in 2012 already had high resistance in 2009 (Figure 1). We Author affiliations: Harvard T.H. Chan School of Public Health, calculated the Spearman correlation coefficient between the Boston, Massachusetts, USA (C.P. Andam, C.J. Worby, proportion of PNSP for each pair of serotypes across states W.P. Hanage); Centers for Disease Control and Prevention, in 2009 (range –0.09 to 0.66) and 2012 (range 0.30–0.79) Atlanta, Georgia, USA (R. Gierke, L. McGee, T. Pilishvili)

1 DOI: https://dx.doi.org/10.3201/eid2306.161331 These first authors contributed equally to this article.

1012 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Penicillin Resistance of Nonvaccine Pneumococcus

Figure 1. Comparison of proportion of nonvaccine type serotypes with penicillin resistance, by serotype, United States, 2009 and 2012. Based on Active Bacterial Core surveillance system data from 10 US states. The dashed diagonal line represents no change.

(online Technical Appendix, https://wwwnc.cdc.gov/EID/ changes are not statistically significant relative to 2009, they article/23/6/16-1331-Techapp1.pdf). We found significant might suggest shifting contributions to the observed varia- overall correlation in 2009 and 2012 (p<0.001 for both tion in proportions of PNSP after the introduction of the years), indicating that sites with high proportions of PNSP in PCV13 vaccine in 2010, with geographic differences in se- 1 serotype typically will also have high proportions of PNSP rotype distribution having an increased role and differences in other serotypes. This finding suggests that differences in in serotype-specific PNSP becoming less important. selection pressure account for the geographic variation in the Finally, we sought to quantify the rate of change in proportions of PNSP. penicillin resistance during 2009–2012 in each state. We We next implemented a standardized regression ap- documented the proportion of PNSP by state for the pre- proach, used previously to analyze the pneumococcal-re- and post-PCV13 periods (online Technical Appendix Ta- sistance patterns pre-PCV7 (7) and post-PCV7 (8) (online ble 4). No significant change in state-level resistance was Technical Appendix). To investigate the source of geo- observed. New Mexico, Maryland, and Georgia saw the graphic variation in the proportion of PNSP, we tested the highest increases in the proportion of PNSP during 2009– hypotheses that either geographic heterogeneity in serotype 2012, whereas a slight decline was observed for Colorado, distribution (Std1), or serotype-specific differences in peni- New York, and Connecticut. Although the distribution of cillin resistance (Std2) were responsible for the observed serotypes might greatly fluctuate among geographic re- variation. These effects were quantified by regressing crude gions immediately after vaccine introduction, the overall versus standardized prevalence of penicillin resistance (Fig- proportions of NVT serotypes with penicillin resistance ure 2; online Technical Appendix), by which a regression across the country might not vary significantly between the slope of 1 would indicate that the factor considered had zero pre- and post-vaccine periods. Of potential importance are effect. By using 2009 data, we found that regression slopes the small increases in the proportions of PNSP serotypes for Std1 (0.445, 95% CI –0.083 to 0.972) and Std2 (0.463, not included in either vaccine that were observed between 95% CI –0.013 to 0.939) indicate that both factors played a the implementation of PCV7 and PCV13 (12), which might similar intermediate role in generating this geographic varia- lay the foundation for changes post-PCV13. tion in penicillin resistance, with neither 95% CI containing 1. In 2012, the regression coefficient for Std2 was higher Conclusions (0.634, 95% CI 0.14–1.128), whereas the coefficient for Std1 The marked variation in the proportion of penicillin decreased (0.367, 95% CI –0.025 to 0.758). Although these resistance among states highlights the potential of local

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1013 DISPATCHES

Figure 2. Crude versus standardized proportions of nonvaccine type serotypes with penicillin nonsusceptibility, by state, United States, 2009 and 2012, based on Active Bacterial Core surveillance system data from 10 US states. Std1 denotes standardization for geographic heterogeneity in serotype distribution. Std2 denotes standardization for serotype-specific differences in resistance. Regression slopes with 95% CIs are indicated in the upper left corner of each panel. Larger circles represent states with a greater number of penicillin-resistant samples. Dashed lines represent the inverse-variance weighted (red) and unweighted (gray) regression slopes. PNSP, penicillin-nonsusceptible pneumococcus.

selective pressures to favor certain serotypes and resistant and Streptococcal Laboratory teams at the Centers for Disease strains within each serotype to increase in frequency as Control and Prevention. the population returns to equilibrium (13). Previous stud- W.P.H. was supported by the National Institute of Allergy and ies have already shown significant regional differences in Infectious Diseases of the National Institutes of Health (award antibiotic use and vaccination coverage across the United no. R01 AI106786-01). C.J.W. was supported by the National States (14,15). Regional rates of patient adherence to treat- Institute of General Medical Sciences of the National Institutes ment regimens will also influence variations in resistance. of Health (award no. U54GM088558). A combination of these factors, which will likely vary between and within regions, would greatly affect proportions The content is solely the responsibility of the authors and of resistance across the country. does not necessarily represent the official views of the In our study, we observed that the dynamics of penicillin National Institutes of Health or the institutions with which the resistance continue to shift in the wake of vaccine introduc- authors are affiliated. The funders had no role in study design, tion. Our postvaccine observations were recorded shortly data collection and analysis, decision to publish, or preparation after the introduction of the vaccine; additional observations of the manuscript. would be valuable to determine the stability of the postvac- Dr. Andam is a postdoctoral fellow at the Department of cine dynamics and any potential importance of the tempo- Epidemiology at the Harvard T.H. Chan School of Public Health. ral changes we observed to factors contributing to variation Her research focuses on using genomic data to understand the in resistance levels. Further long-term nationwide surveil- population structure and evolution of pneumococcus. lance of serotype dynamics is required to assess the multiple ecologic factors that influence antibiotic resistance in the pneumococcus in the conjugate vaccine era. References 1. Hsu KK, Shea KM, Stevenson AE, Pelton SI; Massachusetts Department of Public Health. Changing serotypes causing Acknowledgments childhood invasive pneumococcal disease: Massachusetts, We thank the principal investigators and surveillance officers 2001–2007. Pediatr Infect Dis J. 2010;29:289–93. at the 10 participating ABCs sites and the ABCs Epidemiology https://doi.org/10.1097/INF.0b013e3181c15471

1014 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Penicillin Resistance of Nonvaccine Pneumococcus

2. Scott JR, Millar EV, Lipsitch M, Moulton LH, Weatherholtz R, Perilla MJ, et al. Impact of more than a decade of pneumococcal conjugate vaccine use on carriage and invasive potential in EID Podcast: Native American communities. J Infect Dis. 2012;205:280–8. http://dx.doi.org/10.1093/infdis/jir730 Probable Unusual 3. Huang SS, Platt R, Rifas-Shiman SL, Pelton SI, Goldmann D, Finkelstein JA. Post-PCV7 changes in colonizing pneumococcal Transmission of Zika Virus serotypes in 16 Massachusetts communities, 2001 and 2004. Pediatrics. 2005;116:e408–13. http://dx.doi.org/10.1542/ peds.2004-2338 4. Croucher NJ, Finkelstein JA, Pelton SI, Mitchell PK, Lee GM, Parkhill J, et al. Population genomics of post-vaccine changes in pneumococcal epidemiology. Nat Genet. 2013;45:656–63. http://dx.doi.org/10.1038/ng.2625 5. Weinberger DM, Malley R, Lipsitch M. Serotype replacement in disease after pneumococcal vaccination. Lancet. 2011;378:1962– 73. http://dx.doi.org/10.1016/S0140-6736(10)62225-8 6. Jefferies JMC, Macdonald E, Faust SN, Clarke SC. 13-valent pneumococcal conjugate vaccine (PCV13). Hum Vaccin. 2011;7:1012–8. http://dx.doi.org/10.4161/hv.7.10.16794 7. McCormick AW, Whitney CG, Farley MM, Lynfield R, Harrison LH, Bennett NM, et al. Geographic diversity and temporal trends of antimicrobial resistance in Streptococcus pneumoniae in the United States. Nat Med. 2003;9:424–30. http://dx.doi.org/10.1038/nm839 8. Link-Gelles R, Thomas A, Lynfield R, Petit S, Schaffner W, Harrison L, et al. Geographic and temporal trends in antimicrobial nonsusceptibility in Streptococcus pneumoniae in the post-vaccine era in the United States. J Infect Dis. 2013;208:1266–73. http://dx.doi.org/10.1093/infdis/jit315 9. Hicks LA, Chien Y-W, Taylor TH Jr, Haber M, Klugman KP; Ac- Zika virus (ZIKV), a mosquito-transmitted tive Bacterial Core Surveillance (ABCs) Team. Outpatient flavivirus, has been isolated from sentinel mon- antibiotic prescribing and nonsusceptible Streptococcus pneumoniae in the United States, 1996–2003. Clin Infect Dis. 2011; keys, mosquitoes, and sick persons in Africa and 53:631–9. http://dx.doi.org/10.1093/cid/cir443 Southeast Asia. Serologic surveys indicate that 10. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing; 16th informational ZIKV infections can be relatively common among supplement (M100–S16). Wayne (PA): The Institute; 2008. persons in southeastern Senegal and other areas 11. Venkateswaran PS, Stanton N, Austrian R. Type variation of strains of Africa, but that ZIKV-associated disease may be of Streptococcus pneumoniae in capsular serogroup 15. J Infect Dis. 1983;147:1041–54. http://dx.doi.org/10.1093/infdis/147.6.1041 underreported or misdiagnosed. In 2007, a large 12. Gertz RE Jr, Li Z, Pimenta FC, Jackson D, Juni BA, Lynfield R, outbreak of ZIKV infection occurred on Yap Island et al.; Active Bacterial Core Surveillance Team. Increased in the southwestern Pacific that infected ≈70% penicillin nonsusceptibility of nonvaccine-serotype invasive pneumococci other than serotypes 19A and 6A in post–7-valent of the island’s inhabitants, which highlighted this conjugate vaccine era. J Infect Dis. 2010;201:770–5. virus as an emerging pathogen. The purpose of http://dx.doi.org/10.1086/650496 13. Hanage WP, Bishop CJ, Lee GM, Lipsitch M, Stevenson A, this study was to investigate and report 3 unusual Rifas-Shiman SL, et al. Clonal replacement among 19A cases of arboviral disease that occurred in Colo- Streptococcus pneumoniae in Massachusetts, prior to 13 rado in 2008 valent conjugate vaccination. Vaccine. 2011;29:8877–81. http://dx.doi.org/10.1016/j.vaccine.2011.09.075 Clinical and serologic evidence indicate that 14. Suda KJ, Hicks LA, Roberts RM, Hunkler RJ, Taylor TH. Trends two American scientists contracted Zika virus and seasonal variation in outpatient antibiotic prescription rates in infections while working in Senegal in 2008. One the United States, 2006 to 2010. Antimicrob Agents Chemother. 2014;58:2763–6. http://dx.doi.org/10.1128/AAC.02239-13 of the scientists transmitted this arbovirus to 15. Centers for Disease Control and Prevention. Invasive his wife after his return home. Direct contact is pneumococcal disease and 13-valent pneumococcal conjugate vaccine (PCV13) coverage among children aged ≤59 months— implicated as the transmission route, most likely selected U.S. regions, 2010–2011. MMWR Morb Mortal Wkly as a sexually transmitted infection. Rep. 2011;60:1477–81.

Address for correspondence: Cheryl P. Andam, Harvard University Visit our website to listen: T.H. Chan School of Public Health – Epidemiology, 677 Huntington https://www2c.cdc.gov/ Ave, Boston, MA 02115, USA; email: [email protected] or podcasts/player. [email protected] asp?f=7106489

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1015 DISPATCHES

Febrile Respiratory Illness Associated with Human Adenovirus Type 55 in South Korea Military, 2014–20161

Hongseok Yoo, Se Hun Gu, Jaehun Jung, experienced severe manifestations. We describe the epide- Dong Hyun Song, Changgyo Yoon, miologic, clinical, and molecular characteristics of FRI in Duck Jin Hong, Eun Young Lee, Woong Seog, the South Korea military during October 2014–May 2016. Il-Ung Hwang, Daesang Lee, Seong Tae Jeong,2 Kyungmin Huh2 The Study We obtained data regarding temporal trends in FRI inci- An outbreak of febrile respiratory illness associated with dence from military sentinel surveillance, which has been human adenovirus (HAdV) occurred in the South Korea monitoring weekly FRI rates since October 2011. Monthly military during the 2014–15 influenza season and thereaf- numbers of patients with pneumonia (inpatients, outpa- ter. Molecular typing and phylogenetic analysis of patient samples identified HAdV type 55 as the causative agent. tients, and emergency room patients) were extracted from a Emergence of this novel HAdV necessitates continued sur- computerized data warehouse that stores data from all mili- veillance in military and civilian populations. tary hospitals. We identified pneumonia cases by using the International Classification of Diseases and Related Health Problems, 10th Revision, codes J12–J18. The influenza sea- uman adenovirus (HAdV) is a common cause of son, which starts in October and ends the following May, Hrespiratory infections ranging from uncomplicated was used as a surrogate for the HAdV season in this study. upper respiratory infections to life-threatening pneumo- More detaileds information on FRI surveillance is available nia. Military personnel, especially new recruits, are pre- in the online Technical Appendix (https://wwwnc.cdc.gov/ disposed to respiratory infections caused by HAdV (1). EID/article/23/6/16-1848-Techapp1.pdf). The substantial effects of HAdV infection in the military The trends in FRI rates showed an unusual surge dur- have been demonstrated by the marked increase in the in- ing the 2014–15 influenza season (Figure 1, panel A). The cidence of febrile respiratory illness (FRI) in the US mili- FRI rate increased for 15 weeks in the 2014–15 season, tary after vaccination against the virus ended in 1999; in compared with 10 weeks in the 2012–13 season and 5 turn, the incidence dramatically declined after the vaccine weeks in the 2013–14 season. Peak FRI rate in the 2014–15 was reintroduced (2). season (10.4%) was higher than rates in the preceding 2 HAdVs are a group of nonenveloped double-stranded seasons (4.7% and 7.5%). The numbers of pneumonia cas- DNA viruses comprising 7 species (A–G) and >50 types (3). es in 2014–15 and 2015–16 seasons were 3,140 and 3,145 HAdV types belonging to species B (HAdV-3, -7, -11, -16, patients, respectively, a 191% increase from the mean -21) and E (HAdV-4) are commonly associated with respira- number during 3 preceding seasons. tory infections in adults, particularly in military personnel In April 2014, a multiplex real-time PCR for identify- (4). Novel types or genomic variants, such as HAdV-14 (3) ing 15 viruses from respiratory specimens was introduced and HAdV-7 (5), have been implicated in epidemics of se- at the Armed Forces Capital Hospital, the only tertiary hos- vere infection. HAdV-55, another emerging type reported in pital in the South Korea military health care system (de- China, Turkey, Spain, Singapore, and Israel (5,6), has been tailed methods in online Technical Appendix). A total of associated with severe clinical manifestations, which often 1,484 nonduplicate specimens were tested by the end of lead to respiratory failure and death (7,8). May 2016 (Figure 1, panel B; online Technical Appendix Since fall 2014, we have observed an outbreak of Table 1, Figure). HAdV was identified in 490 (33.0%) of FRI and pneumonia in military personnel in South Korea. total specimens, and it accounted for 79.7% (282/354) and In addition to the increased incidence of FRI, patients 53.2% (150/282) of positive results in the 2014–15 and Author affiliations: Armed Forces Capital Hospital, Seongnam, 2015–16 seasons, respectively. Increased HAdV activity South Korea (H. Yoo, D.J. Hong, E.Y. Lee, K. Huh); Agency for was observed from December until the following May. Defense Development, Daejeon, South Korea (S.H. Gu, 1Preliminary results from this study were presented at the Annual D.H. Song, D. Lee, S.T. Jeong); Armed Forces Medical Command, Spring Meeting of the Korean Society for Chemotherapy, April Seongnam (J. Jung, C. Yoon, W. Seog, I.-U. Hwang) 21–22, Seoul, South Korea. DOI: http://dx.doi.org/10.3201/eid2306.161848 2These authors contributed equally to this article.

1016 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Adenovirus Type 55 in South Korea Military

Figure 1. A) Weekly febrile respiratory illness (FRI) rate (solid line) and monthly number of pneumonia patients (dashed line) in the South Korea military, 2011–2016. B) Positive rate of human adenovirus from respiratory specimens (red line) and the number of respiratory virus PCR requested (blue bar) from a tertiary military hospital, South Korea, 2014–2016. The rate and number for each month are shown in the table at bottom.

We reviewed the demographic and clinical informa- HAdV-infected patients had a significantly higher risk of tion of 878 military patients with FRI or pneumonia who requiring intensive care or mechanical ventilator support. were tested for respiratory viruses from October 2014 In the HAdV-infected group, 8 patients required intubation through May 2016 (Tables 1, 2). Soldiers of lower rank and 1 died; no one in the noninfected group died or required were markedly more likely to infected with HAdV; sol- intubation. Length of hospital stay was also significantly diers serving in the Air Force were less likely. Patients longer among those the HAdV-infected group than among who had been referred from other hospitals were twice as those in the noninfected group (12.6 vs. 9.4 days). likely to be HAdV-infected than patients who visited the We conducted molecular typing by the sequencing of Armed Forces Capital Hospital directly. Rhinorrhea, sore hexon and fiber genes with 74 HAdV-positive respiratory throat, diarrhea, and nausea/vomiting were more common specimens collected from March through June 2016 (meth- in patients with HAdV infection. The proportion of patients ods and general characteristics of the patients are available with pneumonia and the hospitalization rate did not differ in the online Technical Appendix Table 2). Among them, 49 between those with and without HAdV infection. However, samples were successfully sequenced (GenBank numbers in

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1017 DISPATCHES

Table 1. Epidemiologic characteristics of patients with or without identification of HAdV from respiratory specimens by PCR, South Korea, 2014–2016* Patients with HAdV, Patients with other virus PCR Epidemiologic characteristic n = 447 negative for HAdV, n = 431 OR (95% CI) p value Year Apr 2014–May 2015 274 (65.4) 145 (34.6) 3.13 (2.37–4.12) <0.001 Jun 2015–May 2016 173 (37.7) 286 (62.3) Rank Recruit or private 251 (70.5) 105 (29.5) 3.98 (2.98–5.31) <0.001 PFC or higher 196 (37.5) 326 (62.5) Service Army 423 (52.1) 391 (47.9) N\A <0.001 Navy/Marine Corps 20 (50.0) 20 (50.0) Air Force 1 (4.8) 20 (95.2) Region Seoul/Gyeonggi-do 376 (50.0) 376 (50.0) 0.055 Gangwon-do 46 (65.7) 24 (34.3) Chungcheong-do 12 (38.7) 19 (61.3) Gyeongsang-do 10 (58.8) 7 (41.2) Jeolla-do 3 (37.5) 5 (62.5) Route of visit Direct 330 (47.3) 367 (52.7) 2.03 (1.45–2.85) <0.001 Referral 117 (64.6) 64 (35.4) Age, y, mean (SD) 20.8 (2.0) 22.2 (5.0) <0.001 *Values are no. (%) except as indicated. HAdV, human adenovirus; NA, not available; OR, odds ratio; PFC, private first class. online Technical Appendix). Phylogenetic analyses showed The most notable finding of our study is the emergence that all 49 HAdV strains from South Korea clustered with of HAdV-55 in the South Korea military. HAdV-55 is a nov- HAdV-55 strains from China, Singapore, Taiwan, Spain, el type that has been associated with a severe clinical course and the United States (Figure 2). and death in healthy young adults (7,8). We also found that HAdV infection was associated with intensive care, mechan- Conclusions ical ventilator support, and longer hospital stay. In addition, We describe an outbreak of FRI associated with HAdV in we found that the only patient who died was HAdV infected. the South Korea military. HAdV is a well-known major From a molecular perspective, HAdV-55 is a novel type with cause of FRI in the military, accounting for >50% of FRI a hexon gene recombination between HAdV-11 and HAdV- and pneumonia cases in military recruits (1). Our study also 14 (11). Phylogenetic analysis by using the hexon and fiber confirmed the predominance of HAdV, which was identi- gene sequence of 49 strains collected in our study showed that fied in 49.1% of specimens from patients with FRI or pneu- they clustered with previously reported HAdV-55 strains. monia. These findings are similar to those of previous stud- Our findings have implications beyond military set- ies from South Korea and the United States (9,10). tings. Spread of infection of traditionally military-associated

Table 2. Clinical characteristics of patients with or without identification of HAdV from respiratory specimens PCR, South Korea, 2014–2016* Patients with HAdV, Patients with other virus or PCR Clinical characteristic n = 447 negative for HAdV, n = 431 OR (95% CI) p value Presenting symptoms Cough† 423 (94.6) 395 (91.9) 1.56 (0.91–2.67) 0.102 Rhinorrhea‡ 229 (51.3) 192 (44.7) 1.31 (1.00–1.71) 0.047 Sore throat§ 286 (64.3) 207 (48.1) 1.94 (1.48–2.54) <0.001 Dyspnea¶ 60 (13.5) 41 (9.5) 1.48 (0.97–2.25) 0.070 Diarrhea# 125 (33.8) 60 (17.2) 2.46 (1.73–3.49) <0.001 Nausea/vomiting** 115 (31.0) 58 (16.6) 2.25 (1.58–3.22) <0.001 Pneumonia 231 (51.7) 250 (58.0) 0.77 (0.59–1.01) 0.060 Hospitalization 277 (62.0) 270 (62.6) 0.97 (0.74–1.28) 0.836 Intensive care 70 (25.3) 30 (11.1) 2.71 (1.70–4.31) <0.001 Mechanical respiratory support 25 (9.0) 5 (1.9) 5.26 (1.98–13.95) <0.001 Intubation 8 (2.9) 0 NA 0.005 Death 1 (0.4) 0 NA 0.323 Length of stay, d, mean (SD) 12.6 (9.7) 9.4 (5.0) NA <0.001 *Values are no. (%) except as indicated. HAdV, human adenovirus; NA, not available; OR, odds ratio.†n = 877. ‡n = 876. §n = 875. ¶n = 876. #n = 719. **n = 720.

1018 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

Adenovirus Type 55 in South Korea Military

Figure 2. Phylogenetic analysis of human adenoviruses based on the partial nucleotide sequences of hexon (A) and fiber (B) genes, South Korea, 2016. Phylogenetic trees were generated by the neighbor-joining method, using the Kimura 2-parameter method. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1,000 replicates) are shown next to the branches. Red indicates viruses identified in this study. Scale bars denote the number of base substitutions per site.

HAdV types into civilians has been recently reported in the typing study conducted in 2007 reported HAdV-7 as the United States and China (3,12,13). Thus, surveillance of most prevalent type (14). Lack of continuous surveillance HAdV types among both military and civilian populations makes it difficult to estimate exactly when this novel type is warranted; such measures are being implemented by the was introduced into South Korea. US Centers for Disease Control and Prevention (Atlanta, Further genomic analysis of the collected samples and GA, USA) (12). enhanced surveillance, including of civilian populations, Our study has some limitations. First, our findings may would provide more information on the epidemiology of not be generalizable due to the retrospective nature of the HAdV infection. In addition, studies are needed on the ef- study. However, the military health system in South Korea ficacy of previous vaccines against HAdV-55. provides healthcare exclusively to all military personnel; therefore, epidemiologic information gathered from our Acknowledgments surveillance is accurate and comprehensive. Second, we We express our sincere gratitude to the healthcare professionals conducted molecular typing with samples collected from who have been involved in the care of the patients with FRI February 2016, which was substantially later than the on- described in this study. set of the epidemic. However, HAdV-55 had already been This work was supported by a grant from the Agency for Defense identified in a case series from our center during June 2014– Development, Republic of Korea (UE134020ID). May 2015 (8). Because evidence shows that HAdV-55 has been already circulating since early 2014, we believe we Dr. Yoo is a pulmonary physician and medical officer at the can assume that HAdV-55 was the causative agent of the Armed Forces Capital Hospital of South Korea. His research outbreak described in this study. Previously, the HAdV interests include pulmonary infectious diseases.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1019 DISPATCHES

References 1. Lynch JP III, Kajon AE. Adenovirus: epidemiology, global spread June 2016: of novel serotypes, and advances in treatment and prevention. Semin Respir Crit Care Med. 2016;37:586–602. http://dx.doi.org/ 10.1055/s-0036-1584923 Respiratory Diseases 2. Radin JM, Hawksworth AW, Blair PJ, Faix DJ, Raman R, • Debate Regarding Oseltamivir Russell KL, et al. Dramatic decline of respiratory illness among US Use for Seasonal and military recruits after the renewed use of adenovirus vaccines. Clin Pandemic Influenza Infect Dis. 2014;59:962–8. http://dx.doi.org/10.1093/cid/ciu507 • Human Infection with 3. Kajon AE, Lu X, Erdman DD, Louie J, Schnurr D, George KS, Influenza A(H7N9s) Virus during et al. Molecular epidemiology and brief history of emerging 3 Major Epidemic Waves, adenovirus 14-associated respiratory disease in the United States. China, 2013–2015 J Infect Dis. 2010;202:93–103. http://dx.doi.org/10.1086/653083 4. Metzgar D, Osuna M, Kajon AE, Hawksworth AW, Irvine M, • Integration of Genomic and Russell KL. Abrupt emergence of diverse species B adenoviruses at Other Epidemiologic Data to US military recruit training centers. J Infect Dis. 2007;196:1465– Investigate and Control a 73. http://dx.doi.org/10.1086/522970 Cross-Institutional Outbreak of 5. Lu Q-B, Tong Y-G, Wo Y, Wang H-Y, Liu E-M, Gray GC, et al. Streptococcus pyogenes Epidemiology of human adenovirus and molecular characterization • Extended Human-to-Human of human adenovirus 55 in China, 2009-2012. Influenza Other Transmission during a • Heterogeneous and Dynamic Respi Viruses. 2014;8:302–8. http://dx.doi.org/10.1111/irv.12232 Monkeypox Outbreak in Prevalence of Asymptomatic 6. Salama M, Amitai Z, Nutman A, Gottesman-Yekutieli T, the Democratic Republic of Influenza Virus Infections Sherbany H, Drori Y, et al. Outbreak of adenovirus type 55 the Congo • Improved Global Capacity for infection in Israel. J Clin Virol. 2016;78:31–5. http://dx.doi.org/ • Experimental Infection and Influenza Surveillance 10.1016/j.jcv.2016.03.002 Response to Rechallenge • Prevalence of Nontuberculous 7. Gu L, Qu J, Sun B, Yu X, Li H, Cao B. Sustained viremia and high of Alpacas with Middle Mycobacterial Pulmonary viral load in respiratory tract secretions are predictors for death in East Respiratory Syndrome Disease, Germany, 2009–2014 immunocompetent adults with adenovirus pneumonia. PLoS One. Coronavirus 2016;11:e0160777. http://dx.doi.org/10.1371/journal.pone.0160777 • Antibody Response and Disease 8. Yoon H, Jhun BW, Kim SJ, Kim K. Clinical characteristics and Severity in Healthcare Worker factors predicting respiratory failure in adenovirus pneumonia. MERS Survivors Respirology. 2016;21:1243–50. http://dx.doi.org/10.1111/resp.12828 • Epidemiology of Pulmonary 9. Heo JY, Lee JE, Kim HK, Choe KW. Acute lower respiratory Nontuberculous Mycobacterial tract infections in soldiers, South Korea, April 2011–March 2012. Disease, Japan Emerg Infect Dis. 2014;20:875–7. http://dx.doi.org/10.3201/ • Elevated Pertussis Reporting eid2005.131692 in Response to 2011–2012 10. Padin DS, Faix D, Brodine S, Lemus H, Hawksworth A, Putnam S, Outbreak, New York City, et al. Retrospective analysis of demographic and clinical factors New York, USA associated with etiology of febrile respiratory illness among US military basic trainees. BMC Infect Dis. 2014;14:576. • Hemophagocytic http://dx.doi.org/10.1186/s12879-014-0576-2 Lymphohistiocytosis and 11. Walsh MP, Seto J, Jones MS, Chodosh J, Xu W, Seto D. Progressive Disseminated Computational analysis identifies human adenovirus type 55 as a • Use of Population Genetics to Histoplasmosis re-emergent acute respiratory disease pathogen. J Clin Microbiol. Assess the Ecology, Evolution, • Next-Generation Sequencing of 2010;48:991–3. http://dx.doi.org/10.1128/JCM.01694-09 and Population Structure Mycobacterium tuberculosis of Coccidioides 12. Scott MK, Chommanard C, Lu X, Appelgate D, Grenz L, • MERS-CoV Infection of Alpaca Schneider E, et al. Human adenovirus associated with severe • Infection, Replication, and in a Region Where MERS-CoV respiratory infection, Oregon, USA, 2013–2014. Emerg Infect Dis. Transmission of Middle East is Endemic 2016;22:1044–51. http://dx.doi.org/10.3201/eid2206.151898 Respiratory Syndrome 13. Cao B, Huang G-H, Pu Z-H, Qu J-X, Yu X-M, Zhu Z, et al. Coronavirus in Alpacas Emergence of community-acquired adenovirus type 55 as a • Rapid Detection of Polymyxin cause of community-onset pneumonia. Chest. 2014;145:79–86. Resistance in Enterobacteriaceae http://dx.doi.org/10.1378/chest.13-1186 14. Jeon K, Kang CI, Yoon CH, Lee DJ, Kim CH, Chung YS, et al. • Human Adenovirus Associated High isolation rate of adenovirus serotype 7 from South Korean with Severe Respiratory military recruits with mild acute respiratory disease. Eur J Clin Infection, Oregon, 2013–2014 Microbiol Infect Dis. 2007;26:481–3. http://dx.doi.org/10.1007/ s10096-007-0312-6 https://wwwnc.cdc.gov/eid/articles/

Address for correspondence: Kyungmin Huh, Division of Infectious issue/22/6/table-of-contents Diseases and Office of Infection Control, Armed Forces Capital Hospital, Seongnam, South Korea; email: [email protected]; Seong Tae Jeong, The 5th R&D Institute, Agency for Defense Development, Yuseong, P.O. Box 35, Daejeon, 34188, South Korea; email: [email protected]

1020 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 COMMENTARY

Stockpiling Ventilators for Influenza Pandemics

Martin I. Meltzer, Anita Patel

or decades, public health officials have been well aware for such resources. The authors mention that a potential Fof the importance of planning and preparing for the benefit of stockpiling ventilators at hospitals is to facilitate next influenza pandemic (1). An open-access, spreadsheet- staff training; however, only a few (1–3) ventilators would based tool, FluSurge 2.0, enables public health officials and be needed to support training needs. Also, once stockpiles hospital administrators to estimate the potential surge in are established, the costs of replenishing inventory over demand that inevitably will occur during the next influenza time or replacing products to meet changing technology pandemic. The tool provides estimates for hospital-based are not considered. The model also assumes that stockpiled services, such as intensive care unit beds and mechanical ventilators will not be used for noninfluenza patients. In ventilators (2,3). Any influenza pandemic has the potential reality, stockpiled ventilators are likely to be simpler ven- to overwhelm existing hospital capacities, supplies, and tilators that can be used on more stable patients, thus free- readily available means of resupply. Thus, all capable pub- ing up other sophisticated ventilators for patients requiring lic health services should develop and carry out plans to greater respiratory support. stockpile critical resources, such as mechanical ventilators, An important aspect of interpreting the results from needed to support patients who become severely ill from the model of Huang et al. is the problem of expected un- pandemic influenza (4). A major challenge to planning is met demand. There is an upper limit to the number of ad- determining the amount of ventilators to stockpile and conditional ventilators that any hospital can absorb and use to sidering how to manage them so they are effectively used successfully help treat acutely ill patients needing mechani- during the pandemic response. cal ventilation. This limit is determined in large part by the In this issue (5), Huang et al. provide a valuable number of trained staff—particularly respiratory therapists, addition to public health preparedness by presenting a nurses, and technicians—available to ventilate and moni- planning tool to help public health officials consider the tor patients (6). That is, the number of machines is less of a optimal size of a stockpile of mechanical ventilators, constraint than is availability of trained personnel. Huang et compiled on the basis of data from Texas, USA. Users of al. allow for expected unmet needs, setting a default value their tool also can evaluate where such a stockpile should of an acceptable level of 5 patients unable to receive me- be stored: a central location, or prepositioned strategically chanical ventilation at any given time. Assuming a moderate, in hospital facilities. To determine how many ventilators 2009-type influenza pandemic, the authors estimate a 30% to stockpile and where to stockpile them, the authors ex- chance of this expected unmet need occurring. To meet this plicitly include in their tool key factors such as timing level of unmet demand in Texas, planning for a moderate or of a pandemic’s peak locally (not all regions will experi- a severe pandemic requires stockpiling as few as 1,172 or ence simultaneous peak demands for ventilators); wastage as many as 15,697 ventilators, respectively. However, actu- (ventilators not sent to where they are needed, when they ally deploying and using such high volumes of mechanical are most needed, and/or cannot be used); and expected ventilators would be challenging in terms of having enough unmet demand for mechanical ventilators (i.e., when a hospital space and staff to support additional ventilator use. hospital has more patients who need mechanical ventila- Thus, during moderate and severe pandemics, a higher level tion than available ventilators). of unmet demand might need to be expected. Attending phy- All mathematical models have limitations, and some sicians will have to determine who gets access to the lim- important practical problems related to ventilator prepared- ited number of ventilators and who does not. Only a small ness are beyond the reasonable scope of the model by number of studies describe how physicians might make such Huang et al. For example, hospitals must accept respon- allocation decisions for critical, scarce resources (i.e., triage sibility for the costs and resources needed to manage and or prioritization) and how they would explain such decisions maintain an excess of ventilators that are likely to be un- to the patients and their families (7–9). used in the absence of pandemic-related surges in demand Tools of the type produced by Huang et al. are essential to adequately plan and optimize the use of stock- Author affiliation: Centers for Disease Control and Prevention, piled resources during the next influenza pandemic. Such Atlanta, Georgia, USA tools, however, are just one part of systematic planning, DOI: https://dx.doi.org/10.3201/eid2306.170434 which should include elements such as the number to be

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1021 COMMENTARY stockpiled, how and where stockpiles should be held, main- 3. Centers for Disease Control and Prevention. FluSurge, v. 2.0 tenance of stockpiles, conditions for release, considerations [cited 2017 Feb 20]. https://www.cdc.gov/flu/pandemic-resources/ tools/flusurge.htm for use, and what to do when stockpiles are insufficient to 4. US Department of Health and Human Services. HHS pandemic adequately meet surges in demand. influenza plan. November 2005 [cited 2017 Feb 20]. https://www. cdc.gov/flu/pandemic-resources/pdf/hhspandemicinfluenzaplan.pdf Dr. Meltzer is an Associate Editor of Emerging Infectious 5. Huang HC, Araz OM, Morton DP, Johnson GP, Damien P, Diseases and leads the Health Economics and Modeling Unit Clements B, et al. Stockpiling ventilators for influenza pandemics. in CDC’s Division of Preparedness and Emerging Infections, Emerg Infect Dis. 2017;23:923–30. 6. Ajao A, Nystrom SV, Koonin LM, Patel A, Howell DR, Baccam P, National Center for Emerging and Zoonotic Infectious Diseases. et al. Assessing the capacity of the US health care system to use His research interests include estimating the impact of influenza additional mechanical ventilators during a large-scale public health pandemics and the economics of controlling infectious diseases. emergency. Disaster Med Public Health Prep. 2015;9:634–41. http://dx.doi.org/10.1017/dmp.2015.105 Dr. Patel is a senior advisor and the lead for pandemic medical care 7. Christian MD, Hawryluck L, Wax RS, Cook T, Lazar NM, and countermeasures with CDC’s Influenza Coordination Unit, Herridge MS, et al. Development of a triage protocol for critical care during an influenza pandemic. CMAJ. 2006;175:1377–81. Division of Viral Diseases, National Center for Immunization and 8. Powell T, Christ KC, Birkhead GS. Allocation of ventilators in a Respiratory Diseases. Her research interests include developing public health disaster. Disaster Med Public Health Prep. 2008;2:20– and implementing science-based operational solutions for medical 6. http://dx.doi.org/10.1097/DMP.0b013e3181620794 countermeasure needed to support a public health response. 9. Ventilator Document Workgroup, Ethics Subcommittee of the Advisory Committee to the Director, Centers for Disease Control and Prevention. Ethical considerations for decision making regarding allocation of mechanical ventilators during a severe References influenza pandemic or other public health emergency [cited 2017 1. Meltzer MI, Cox NJ, Fukuda K. The economic impact of pandemic Mar 16]. https://www.cdc.gov/about/pdf/advisory/ventdocument_ influenza in the United States: priorities for intervention. Emerg release.pdf Infect Dis. 1999;5:659–71. http://dx.doi.org/10.3201/eid0505.990507 2. Zhang X, Meltzer MI, Wortley PM. FluSurge—a tool to estimate demand for hospital services during the next pandemic influenza. Address for correspondence: Martin I. Meltzer, Centers for Disease Med Decis Making. 2006;26:617–23. http://dx.doi.org/10.1177/ Control and Prevention, 1600 Clifton Rd NE, Mailstop C18, Atlanta, GA 0272989X06295359 30329-4027, USA; email: [email protected]

The Public Health Image Library (PHIL) The Public Health Image Library (PHIL), Centers for Disease Control and Prevention, contains thousands of public health-related images, including high-resolution (print quality) photographs, illustrations, and videos.

PHIL collections illustrate current events and articles, supply visual content for health promotion brochures, document the effects of disease, and enhance instructional media.

PHIL images, accessible to PC and Macintosh users, are in the public domain and available without charge.

Visit PHIL at http://phil.cdc.gov/phil

1022 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS

Epidemiologic Survey of mosquitoes. To detect JEV, we used the TIANamp Vi- rus DNA/RNA Kit (TianGen, Beijing, China) with pools Japanese Encephalitis Virus of whole-body Culex mosquito extracts; we analyzed the Infection, Tibet, China, 2015 samples by reverse transcription PCR (RT-PCR) amplification using the Quant One-Step RT-PCR Kit (TianGen) and primers specific for the JEV nonstructural 1 gene (6). Of 11 Hui Zhang, Mujeeb Ur Rehman, Kun Li, C. tritaeniorhynchus and 69 C. pipiens mosquito pools, 7 Houqiang Luo, Yanfang Lan, Fazul Nabi, (63.6%) and 2 (2.9%), respectively, were positive for JEV Lihong Zhang, Muhammad Kashif Iqbal, by RT-PCR; the 1 C. bitaeniorhynchus mosquito pool was Suolangsi Zhu, Muhammad Tariq Javed, not positive for JEV (Table). Yangzom Chamba, Jia Kui Li To determine the origin of JEV in pigs, we collected Author affiliations: Huazhong Agricultural University College of a total of 454 serum samples from 1- to 6-month-old pigs Veterinary Medicine, Wuhan, China (H. Zhang, M.U. Rehman, from local slaughterhouses. We analyzed the samples for K. Li, H. Luo, Y. Lan, F. Nabi, L. Zhang, M.K. Iqbal, J.K. Li); JEV IgM by using a commercial ELISA kit as previously 2 Tibet Agriculture and Animal Husbandry College, Linzhi, Tibet, described (7). We used the χ test and SPSS software (SPSS China (S. Zhu, Y. Chamba, J.K. Li); University of Agriculture, Inc., Chicago, IL, USA) to analyze all data; p<0.05 was Faisalabad, Pakistan (M.T. Javed) considered significant. The overall seroprevalence of JEV IgM in the pigs DOI: https://dx.doi.org/10.3201/eid2306.152115 was 5.07%. The percentage of positive samples from Ny- ingchi County (3.25%) was significantly lower than that We investigated Japanese encephalitis virus (JEV) preva- from Mainling Country (7.81%); no serum samples from lence in high-altitude regions of Tibet, China, by using stan- Gongbo’gyamda County were JEV-positive (Table). The dard assays to test mosquitoes, pigs, and humans. Results difference in seroprevalence of JEV in male (4.62%) and confirmed that JEV has spread to these areas. Disease female (5.67%) pigs was not statistically significant. prevention and control strategies should be used along with surveillance to limit spread of JEV in high-altitude To determine the prevalence of JEV infection among regions of Tibet. persons, we collected blood samples from 364 healthy human volunteers residing in the 3 counties and analyzed the samples for JEV IgM by using a commercial apanese encephalitis virus (JEV) is the causative agent ELISA kit as previously described (8). JEV seropreva- of a viral encephalitis that is a major public health threat lence was11.71% for the 1- to 23-year-old age group, Jin most parts of East and Southeast Asia (1,2). Tibet, China, 13.43% for the 24- to 45-year-old age group, and 4.20% has been considered a nonepidemic area for JEV because for the >45-year-old age group (Table). Seroprevalence its mean elevation is >3,100 m, and the relatively low tem- was significantly higher among persons 1–23 or 24–45 peratures at the elevation do not support JEV transmission years of age, compared with persons >45 years of age, between mosquitoes, reservoirs, and amplifying hosts (2). suggesting that 1) younger persons may have greater ex- However, this situation in Tibet has probably changed due posure risks than persons >45 years of age, 2) JEV is a to increased pig farming in the region and increased travel relative new introduction in the area as a result of the between Tibet and areas where JEV is endemic (3–5). To changing (i.e., warmer) climate, and 3) younger persons test our hypothesis, we tested for JEV in mosquitoes and may travel more frequently than older persons to lower measured JEV-specific IgM in swine and humans living in elevations where JEV is endemic. An IgG-based survey a high-altitude region of Tibet. might identify more JEV disease, even in persons who We conducted the study in Nyingchi District (elevation showed no symptoms of the disease. 3,100 m) in southeastern Tibet. The mean temperature dur- JEV seroprevalence was significantly higher in ru- ing the study period ranged from 11.3°C to 22.4°C. Swine ral populations (6.87%) compared with urban population and humans included in the study had no history of travel (3.02%). The spread of JEV may be increased by amplify- to JEV-endemic areas. ing hosts (pigs); thus, the lower prevalence of JEV in urban To determine whether mosquitoes were infected with residents may be associated with a lower number of pig JEV, we collected 8,330 mosquitoes (belonging to 9 gen- farms in urban areas compared with rural villages in Ny- era and 4 species) near pig sties and human residences in ingchi District. Nyingchi, Mainling, and Gongbo’gyamda Counties. From In conclusion, we found that JEV infection is prevalent those 8,330 mosquitoes, we chose 2,655 JEV vector mos- in a high-altitude region of Tibet that was previously con- quitoes: 330 (3.96%) Culex tritaeniorhynchus, 45 (0.54%) sidered to be free from JEV. Factors such as increased tour- C. bitaeniorhynchus, and 2,280 (27.37%) C. pipiens ism (9), increased mean summer temperatures, increased

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1023 RESEARCH LETTERS

Table. Japanese encephalitis virus IgM–positive pigs and humans and Japanese encephalitis virus-positive Culex mosquitoes in Nyingchi District, Tibet, China, 2015* No. IgM positive/no. tested (%) Total no. positive/total no. tested Category, variable Male pig or person Female pig or person (%) Pigs County location Nyingchi 3/146 (2.05) 5/100 (5.00) 8/246 (3.25) Mainling 9/105 (8.57) 6/87 (6.90) 15/192 (7.81) Gongbo'gyamda 0/9 0/7 0/16 Total 12/260 (4.62) 11/194 (5.67) 23/454 (5.07) Humans County location Nyingchi 6/66 (9.09) 11/93 (11.83) 17/159 (10.69) Mainling 8/58 (13.79) 9/65 (13.85) 17/123 (13.82) Gongbo’gyamda 1/43 (2.33) 1/39 (2.56) 2/82 (2.44) Total 15/167 (8.98) 21/197 (10.66) 36/364 (9.89) Age group, y 1–23 6/57 (10.53) 7/54 (12.96) 13/111 (11.71) 24–45 7/54 (12.96) 11/80 (13.75) 18/134 (13.43) >45 2/56 (3.57) 3/63 (4.76) 5/119 (4.20) Total 15/167 (8.98) 21/197 (10.66) 36/364 (9.89) Culex mosquitoes† C. tritaeniorhynchus ND ND 7/11 (0.636)‡ C. bitaeniorhynchus ND ND 0/1‡ C. pipiens ND ND 2/69 (0.029)‡ *Except as indicated, data are no. (%). ND, Not done. †Denominator indicates no. in pool. ‡Maximum likelihood estimates (no. pools positive for JEV by reverse transcription PCR). movement between the study area and nearby JEV- 4. National Bureau of Statistics of the People’s Republic of China. endemic regions, inadequate public health systems, in- National data. Number of hogs at year-end in Tibet from 1995 to 2014 [cited 2015 Dec 29]. http://data.stats.gov.cn/english/ creased pig farming, increased migration of water birds, easyquery.htm and the absence of a compulsory immunization policy 5. Liu Q, Cao L, Zhu XQ. Major emerging and re-emerging may contribute to emergence of this disease (3,4,9,10). zoonoses in China: a matter of global health and socioeconomic Disease prevention and control strategies should be used development for 1.3 billion. Int J Infect Dis. 2014;25:65–72. http://dx.doi.org/10.1016/j.ijid.2014.04.003 along with surveillance to limit the spread of JEV in high- 6. Hua RH, Liu LK, Chen ZS, Li YN, Bu ZG. Comprehensive altitude regions of Tibet. mapping antigenic epitopes of NS1 protein of Japanese encephalitis virus with monoclonal antibodies. PLoS One. This study was supported by the Key Science Fund of the 2013;8:e67553. http://dx.doi.org/10.1371/journal.pone.0067553 Science and Technology Agency of Tibet Autonomous Region, 7. Pant GR, Lunt RA, Rootes CL, Daniels PW. Serological evidence the Twelfth Five-Year National Science and Technology Support for Japanese encephalitis and West Nile viruses in domestic Project (grant no. 2012BAD3B03), and the Sustentation Fund animals of Nepal. Comp Immunol Microbiol Infect Dis. 2006;29:166–75. http://dx.doi.org/10.1016/j.cimid.2006.03.003 (the serology of Toxoplasma gondii in Tibetan pigs) of the 8. Moore CE, Blacksell SD, Taojaikong T, Jarman RG, Gibbons RV, Huazhong Agricultural University (grant no. 52209-814121). Lee SJ, et al. A prospective assessment of the accuracy of commercial IgM ELISAs in diagnosis of Japanese encephalitis Mr. Zhang is a PhD scholar in the Department of Clinical virus infections in patients with suspected central nervous Veterinary Medicine, Huazhong Agricultural University, Wuhan, system infections in Laos. Am J Trop Med Hyg. 2012;87:171–8. China. His research has focused on the surveillance of emerging http://dx.doi.org/10.4269/ajtmh.2012.11-0729 9. Cao L, Fu S, Gao X, Li M, Cui S, Li X, et al. Low protective infectious diseases by molecular epidemiology analysis and the efficacy of the current Japanese encephalitis vaccine against the analysis of pathogenicity mechanisms. emerging genotype 5 Japanese encephalitis virus. PLoS Negl Trop Dis. 2016;10:e0004686. http://dx.doi.org/10.1371/ References journal.pntd.0004686 1. Misra UK, Kalita J. Overview: Japanese encephalitis. 10. Prosser DJ, Cui P, Takekawa JY, Tang M, Hou Y, Collins BM, Prog Neurobiol. 2010;91:108–20. http://dx.doi.org/10.1016/ et al. Wild bird migration across the Qinghai–Tibetan plateau: j.pneurobio.2010.01.008 a transmission route for highly pathogenic H5N1. PLoS One. 2. Centers for Disease Control and Prevention (CDC). Japanese 2011;6:e17622. http://dx.doi.org/10.1371/journal.pone.0017622 encephalitis surveillance and immunization—Asia and the Western Pacific, 2012. MMWR Morb Mortal Wkly Rep. 2013;62:658–62. Address for correspondence: Jia Kui Li, Veterinary Clinical Medicine, 3. Li YX, Li MH, Fu SH, Chen WX, Liu QY, Zhang HL, et al. College of Veterinary Medicine, Huazhong Agricultural University, Japanese encephalitis, Tibet, China. Emerg Infect Dis. 2011;17:934–6. http://dx.doi.org/10.3201/eid1705.101417 Wuhan, Hubei 430070, China; email: [email protected]

1024 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

RESEARCH LETTERS

High Frequency of Mayaro of CHIKV-negative serum collected during December 2014–June 2015 was tested by hemagglutination inhibition Virus IgM among Febrile (HI) (7) against the most common arboviruses. We evaluat- Patients, Central Brazil ed serum samples with monotypical or heterotypic reactivity to alphaviruses, with HI titers >40 using an IgM MAC ELISA for MAYV, as described by Kuno (8) and then mod- Sandra Brunini, Divânia Dias Silva França, ified (9). The results were sent to the Center for Strategic Juliana Brasiel Silva, Leandro Nascimento Silva, Information in Health Surveillance for further clinical and Flúvia Pereira Amorim Silva, Mariana Spadoni, epidemiologic investigations. Giovanni Rezza We tested 75 samples from DENV-negative patients Author affiliations: Universidade Federal de Goiás, Goiânia, for CHIKV. Five of the 31 DENV-negative persons from Brazil (S. Brunini); Secretaria Municipal de Saúde, Goiânia whom serum was collected during June–December 2014 (D.D.S. França, J.B. Silva, L.N. Silva, F.P.A. Silva); Universidade and none of the 44 from January–June 2015 were positive Federal do Paraná, Paranà, Brazil (M. Spadoni); Istituto Superiore for CHIKV IgM. Two additional samples yielded indeter- di Sanità, Rome, Italy (G. Rezza) minate results for CHIKV. Of 27 CHIKV-negative samples tested by HI, 16 DOI: https://dx.doi.org/10.3201/eid2306.160929 were reactive to alphaviruses (median titer 160 [range 80–1,280]); of these, 15 (56%) were confirmed as positive Mayaro virus (MAYV), an Aedes mosquito–borne alphavirus, by IgM MAC ELISA for MAYV, and 1 sample was bor- is endemic to Brazil and other South America countries. We derline. The median interval between symptom onset and investigated dengue- and chikungunya-negative febrile pa- serum collection was 37 days (range 12–167 days). tients visiting rural areas near Goiânia, Goiás, and found a Patients were a median of 45 years of age (range 30–70 high proportion (55%) of MAYV IgM. Our findings suggest the presence of highly endemic foci of MAYV in central Brazil. years). Fourteen were women. Patients’ educational level was high; 11 (73%) patients had >8 years’ education, and 47% had >12 years’ education. ayaro virus (MAYV) is an Aedes mosquito–borne al- All the patients with an antibody profile suggesting phavirus of the New World, transmitted mainly by recent MAYV infection resided in Goiânia but had trav- Mtree canopy–dwelling Haemagogus spp. mosquitoes, was eled to rural areas in the 15 days before symptom onset. discovered in Trinidad in 1954. MAYV causes a dengue- Twelve patients had visited farms or small holdings in the like acute febrile illness with arthralgic manifestations (1). forests around towns located 34 km (Hidrolandia) and 46 Exposure to MAYV has been reported in several countries km (Bela Vista) from Goiânia (online Technical Appendix of Central and South America, and Mayaro fever has been Figure, https://wwwnc.cdc.gov/EID/article/23/6/16-0929- identified in French Guiana, Suriname, Venezuela, Peru, Bo- Techapp1.pdf). In particular, 7 patients had been around livia, and Brazil (2,3). In Brazil, Mayaro fever has been re- Bela Vista and 5 in the area of Hidrolandia, whereas only 1 ported in the Amazon region (2,4) and in Mato Grosso (5,6). had been near Pontalina and 2 in Itacaja (Tocantins), which At the end of 2014, after chikungunya virus (CHIKV) is outside the state of Goiás. Ten of these 11 patients re- spread in South America, the Brazil Ministry of Health ported engaging in recreational activities. enhanced the surveillance system for dengue-like illness. The high frequency of MAYV IgM detection among Accordingly, febrile patients attending primary care cen- febrile patients in Goiânia is surprising. Identification of ters are tested for dengue virus (DENV) infection by using recent infections most likely acquired in rural areas and viral antigen (NS1), virus isolation, or reverse transcription forests around the city of Goiânia indicates the existence PCR (up to 5 days after symptom onset), and/or ELISA for of active foci where a sylvatic cycle of MAYV is estab- IgM (after the sixth day). DENV-negative patients fitting lished. The MAYV belt around Goiânia, where epizoot- the clinical and epidemiologic criteria for chikungunya set ics of jungle yellow fever were also reported (http://www. by the Ministry of Health (http://www.saude.gov.br) are paho.org/hq/index.php?option=com_docman&task=doc_ notified to the Center for Strategic Information in Health view&Itemid=270&gid=34247&lang=en), is represented Surveillance of Goiânia city for further investigation. by a wet area hosting small primates, which may play a Blood samples collected during June 1, 2014–June 30, role as virus amplifier. 2015, were stored at –20°C and sent to the Department of Our conclusions are subject to several limitations. Arbovirology and Hemorrhagic Fevers, Instituto Evandro First, selection bias could affect findings regarding the Chagas (Belém, Brazil), to be evaluated for CHIKV by an proportion of MAYV recent infections because of the lack IgM-capture ELISA (MAC ELISA; Centers for Disease of systematic testing for dengue-like illness. Second, we Control and Prevention, Atlanta, GA, USA). A subsample did not conduct PCR to identify acute infections because

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1025 RESEARCH LETTERS

Table. Clinical characteristics of 15 patients positive for IgM 6. Pauvolid-Corrêa A, Juliano RS, Campos Z, Velez J, Nogueira RM, against Mayaro virus, Goiânia, Goiás, Brazil, June 2014– Komar N. Neutralising antibodies for Mayaro virus in Pamtanal, June 2015 Brazil. Mem Inst Oswaldo Cruz. 2015;110:125–33. Sign or symptom No. (%) patients 7. Shope RE. The use of a microhemagglutination-inhibition test to Fever 15 (100) follow antibody response after arthropod-borne virus infection in a Arthralgia 14 (93) community of forest animals. An Microbiol (Rio J). 1963;11 Joint edema 14 (93) (Pt A):167–71. Rash 14 (93) 8. Kuno G, Gómez I, Gubler DJ. Detecting artificial anti-dengue IgM Headache 13 (87) immune complexes using an enzyme-linked immunosorbent assay. Weakness 13 (87) Am J Trop Med Hyg. 1987;36:153–9. Myalgia 12 (80) 9. Vasconcelos PFC, Travassos Da Rosa JF, Guerreiro SC, Eye pain 8 (53) Dégallier N, Travassos Da Rosa ES, Travassos Da Rosa AP. Icterus 4 (27) First register of an epidemic caused by Oropouche virus in the Photophobia 3 (20) states of Maranhão and Goiás, Brazil [in Portuguese]. Rev Inst Severe itching 3 (20) Med Trop Sao Paulo. 1989;31:271–8. http://dx.doi.org/10.1590/ Lymphadenopathy 2 (13) S0036-46651989000400011 Vomiting 2 (13) 10. Figueiredo LTM, Nogueira RMR, Cavalcanti SMB, Schatzmayr H, da Rosa AT. Study of two different enzyme immunoassays for the the samples were collected after the viremic phase. Third, detection of Mayaro virus antibodies. Mem Inst Oswaldo Cruz. plaque-reduction neutralization testing was not performed, 1989;84:303–7. http://dx.doi.org/10.1590/S0074-02761989000300003 and because of the lack of convalescent serum, IgG seroconversion or titer increase were not evaluated; however, Address for correspondence: Giovanni Rezza, Department of Infectious MAC ELISA is considered a valid technique for diagnos- Diseases, Istituto Superiore di Sanità, Viale Regina Elena 161, 00142 ing recently acquired infection with MAYV (10). Finally, Rome, Italy; email: [email protected] the frequency of rash (Table), higher than in other case series (4), might be overestimated because of stringent selection criteria used for MAYV testing. In conclusion, infection with MAYV occurs more frequently than expected in central Brazil. Mayaro fever should be considered in the differential diagnosis with DENV, CHIKV, and Zika virus infections in areas charac- Ebola Virus Imported from terized by arbovirus cocirculation. Guinea to Senegal, 2014 Acknowledgment We thank Lívia Carício Martins for supporting testing activities. Daye Ka,1 Gamou Fall,1 Viviane Cissé Diallo, Dr. Brunini is an associate professor at the Faculty of Nursing, Ousmane Faye, Louise Deguenonvo Fortes, Federal University of Goiás. Her primary research interests Oumar Faye, Elhadji Ibrahim Bah, include infectious disease epidemiology and HIV and sexually Kadia Mbaye Diallo, Fanny Balique, transmitted diseases. Cheikh Tidiane Ndour, Moussa Seydi,2 Amadou Alpha Sall2 References Author affiliations: Centre Hospitalier Universitaire de Fann, 1. Anderson CR, Downs WG, Wattley GH, Ahin NW, Reese AA. Dakar, Senegal (D. Ka, V. Cissé Diallo, L. Deguenonvo Fortes, Mayaro virus: a new human disease agent. II. Isolation from blood of patients in Trinidad, B.W.I. Am J Trop Med Hyg. K.M. Diallo, C.T. Ndour, M. Seydi); Institut Pasteur de Dakar, 1957;6:1012–6. Dakar (G. Fall, Ousmane Faye, Oumar Faye, F. Balique, 2. Azevedo RSS, Silva EVP, Carvalho VL, Rodrigues SG, A.A. Sall); Hôpital National Donka, Conakry, Guinea (E.I. Bah) Nunes-Neto JP, Monteiro H, et al. Mayaro fever virus, Brazilian Amazon. Emerg Infect Dis. 2009;15:1830–2. http://dx.doi.org/ DOI: https://dx.doi.org/10.3201/eid2306.161092 10.3201/eid1511.090461 3. Figueiredo ML, Figueiredo LTM. Emerging alphaviruses in the Americas. Rev Soc Bras Med Trop. 2014;47:677–83. In March 2014, the World Health Organization declared an http://dx.doi.org/10.1590/0037-8682-0246-2014 outbreak of Ebola virus disease in Guinea. In August 2014, 4. Mourão MP, Bastos Mde S, de Figueiredo RP, Gimaque JB, a case caused by virus imported from Guinea occurred in Galusso Edos S, Kramer VM, et al. Mayaro fever in the city of Senegal, most likely resulting from nonsecure funerals and Manaus, Brazil, 2007–2008. Vector Borne Zoonotic Dis. 2012:42–6. travel. Preparedness and surveillance in Senegal probably http://dx.doi.org/10.1089/vbz.2011.0669 prevented secondary cases. 5. Zuchi N, Heinen LB, Santos MA, Pereira FC, Slhessarenko RD. Molecular detection of Mayaro virus during a dengue outbreak in 1 the state of Mato Grosso, central-west Brazil. Mem Inst Oswaldo These authors contributed equally to this article. Cruz. 2014;109:820–3. http://dx.doi.org/10.1590/0074-0276140108 2These authors contributed equally to this article.

1026 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

RESEARCH LETTERS

bola virus disease (EVD) is a hemorrhagic fever caused On August 18, he visited a suburban medical center in the Eby Ebola virus (EBOV); the mortality rate is high (1,2). suburbs of Dakar, where he received treatment for malaria: EBOV was discovered in 1976, simultaneously in Zaire quinine, antipyretic and antimicrobial medications, and in- (now the Democratic Republic of the Congo) and Sudan travenous rehydration. Diarrhea and vomiting stopped on (3,4). Since then, small to large outbreaks have occurred day 4 after illness onset, but fever and asthenia persisted. sporadically in the Democratic Republic of the Congo, Su- On August 26, the patient was admitted to Fann Hospital, dan, Gabon, Uganda, Côte d’Ivoire, and Congo (5–7). Dakar, with slight dehydration, fever (39.2°C), and herpet- In March 2014, the World Health Organization (WHO) ic lesions. Because no epidemiologic link with EVD was reported an EVD outbreak caused by Zaire EBOV in Guin- established, the patient was not isolated. ea (8,9). The main feature of this outbreak was its exten- On August 27, a total of 12 members of the patient’s sion into urban areas and neighboring countries (Liberia, family, all suspected of having EVD, were admitted to an Sierra Leone, Nigeria, Senegal, Mali). Ten countries on 3 Ebola treatment center in Conakry, Guinea; test results in- continents were affected; 28,646 confirmed, probable, and dicated EBOV positivity for 6. Epidemiologic investigation suspected cases and 11,323 deaths were recorded. indicated that a member of this family had traveled to Dakar In August 2014, Senegal was the fifth country in Africa and was hospitalized. The Epidemic Management Commit- to be affected by imported EBOV (10). We described this tee set up by WHO in Guinea established an epidemiologic case, the patient’s itinerary and epidemiologic links with link between the patient in Fann Hospital and the confirmed confirmed case-patients in Guinea, and the evolution of the case-patients in Guinea and quickly informed the health au- disease and the virus. thorities in Senegal. The patient in Fann Hospital finally ac- The patient was a 21-year-old man from Forecariah, knowledged that he had attended his uncle’s nonsecure funer- Guinea, who had traveled by land to Senegal during the al in Guinea on August 10, before coming to Dakar (Figure). night of August 13–14, 2014. The date of his illness on- On August 28, the man was transferred to an isolation set was August 16, 2014; symptoms were fever, vomiting, center, and blood samples were sent to the Institut Pas- diarrhea, yellow or black feces, anorexia, and asthenia. teur laboratory in Dakar, a WHO-approved collaborating

Figure. Timeline for case of Ebola virus disease imported into Senegal from Guinea, 2014. Flags indicate patient information; arrows indicate public health actions. BS, blood sample; RT-PCR, reverse transcription PCR.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1027 RESEARCH LETTERS

Centre for EBOV diagnostics. Real-time reverse transcrip- Her research interests include arbovirus–vector interactions, tion PCR (RT-PCR) was positive for Zaire EBOV; viral mechanisms of arbovirus transmission, and public health load was 2.04 × 104 genome copies/mL. ELISA of the same activities such as diagnosis of arboviruses and hemorrhagic sample detected Zaire EBOV–specific IgM (titer 1:400) fever viruses. and IgG (titer 1:3,200). This case of EVD in Senegal was reported to WHO on August 29. The patient received sup- References portive care, and his clinical course progressed well; on 1. Cenciarelli O, Pietropaoli S, Malizia A, Carestia M, D’Amico F, August 31, he was apyrexic and his asthenia had decreased. Sassolini A, et al. Ebola virus disease 2013–2014 outbreak in west In terms of virus evolution, a second blood sample tested Africa: an analysis of the epidemic spread and response. Int J Microbiol. 2015;2015:769121. http://dx.doi.org/10.1155/2015/769121 on day 18 after illness onset showed diminution of viral 2. Feldmann H, Klenk HD. Filoviruses. In: S. Baron, editor. Medical 3 load (4.96 × 10 genome copies/mL) and an IgG titer in- Microbiology, 4th ed. Galveston (TX): University of Texas Medical crease to 1:6,400. A third blood sample collected on day Branch; 1996. 20 showed a negative RT-PCR result, but a urine sample 3. World Health Organization. Report of an International Commission: Ebola haemorrhagic fever in Zaïre, 1976. Bull World collected the same day showed a positive result with a viral Health Organ. 1978;56:271e93. 4 load of 2.04 × 10 genome copies/mL. RT-PCRs of blood 4. World Health Organization/International Study Team. Ebola and urine collected on days 24 and 34 were negative, and haemorrhagic fever in Sudan, 1976. Report of a WHO/International serologic analyses showed a high IgG titer (1:12,800). Study Team. Bull World Health Organ. 1978;56:247–70. 5. Centers for Disease Control and Prevention. Outbreaks chronology: The patient was declared cured on September 18, 2014. Ebola virus disease [cited 2015 Aug 2]. http://www.cdc.gov/vhf/ Epidemiologic investigations revealed a total of 74 contacts ebola/outbreaks/history/chronology.html in Senegal, including 41 healthcare workers (from the sub- 6. World Health Organization. Ebola virus disease [cited 2015 Nov urban medical center and Fann Hospital). Symptoms devel- 28]. http://www.who.int/mediacentre/factsheets/fs103/en/ 7. Colebunders R, Borchert M. Ebola haemorrhagic fever—a review. oped in 5 of these contacts, but their test results were negative J Infect. 2000;40:16–20. http://dx.doi.org/10.1053/jinf.1999.0603 for EBOV. No secondary case was detected after 42 days of 8. World Health Organization. Ebola virus disease in Guinea [cited monitoring, and the outbreak in Senegal was declared over 2015 Aug 18]. http://www.who.int/csr/don/2014_03_23_ebola/en on October 17, 2014, with only 1 confirmed case reported. 9. Baize S, Pannetier D, Oestereich L, Rieger T, Koivogui L, Magassouba N, et al. Emergence of Zaire Ebola virus disease The case-patient’s low viral load, detected during the in Guinea. N Engl J Med. 2014;371:1418–25. http://dx.doi.org/ first RT-PCR 10 days after illness onset, probably explains 10.1056/NEJMoa1404505 the absence of secondary cases in Fann Hospital. However, 10. World Health Organization. Ebola virus disease update— the absence of secondary cases in the suburban medical Senegal [cited 2014 Nov 15] http://www.who.int/csr/don/ 2014_08_30_ebola/en/ center that the patient had visited on days 3–4 after illness onset and among the family members in Dakar is a rare Address for correspondence: Ousmane Faye, Virology Pole, Institut feature of EVD. The preparedness and surveillance estab- Pasteur de Dakar, BP 220 Dakar, Senegal; email: [email protected] lished in Senegal after announcement of EVD in Guinea led to training of healthcare workers for proper use of protective equipment and security procedures with any patient, which probably prevented virus spread in the suburban medical center. This case of EBOV importation from Guinea to Senegal confirms the problems encountered with Ebola outbreak management, including the roles of nonsecure funerals and travel in virus spread. Tick-Borne Encephalitis Virus in Ticks and Roe Deer, Acknowledgments the Netherlands We thank Moussa Dia, El Hadji Abdourahmane Faye, Ousmane Kébé, Khardiata Mbaye, Davy Evrard Kiori, and Oumar Ndiaye Setareh Jahfari, Ankje de Vries, Jolianne M. Rijks, for their excellent technical assistance in laboratory diagnosis. Steven Van Gucht, Harry Vennema, Hein Sprong, This work was supported by grants from the Institut Pasteur de Barry Rockx Dakar, Senegal, and the Ministry of Health, Senegal. Author affiliations: National Institute for Public Health and the Environment, Bilthoven, the Netherlands (S. Jahfari, A. de Vries, Dr. Ka is an infectious disease physician who works in the H. Vennema, H. Sprong, B. Rockx); Utrecht University, Utrecht, Infectious and Tropical Diseases Clinic, Fann Hospital, Dakar, the Netherlands (J.M. Rijks); Scientific Institute of Public Health, Senegal. His research interests are EVD, HIV, and hepatitis. Brussels, Belgium (S. Van Gucht) Dr. Fall is a virologist who works at Arbovirus and Viral Hemorrhagic Fever Unit, Institut Pasteur de Dakar, Senegal. DOI: https://dx.doi.org/10.3201/eid2306.161247

1028 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS

We report the presence of tick-borne encephalitis virus deer serum samples. Serologic screening of all 297 samples (TBEV) in the Netherlands. Serologic screening of roe by ELISA yielded 6 positive and 8 borderline results. All deer found TBEV-neutralizing antibodies with a seropreva- positive, 7 borderline, and 3 negative serum samples were lence of 2%, and TBEV RNA was detected in 2 ticks from confirmed by testing in a TBEV serum neutralization test the same location. Enhanced surveillance and awareness (SNT), with the Neudörfl strain as the accepted prototype among medical professionals has led to the identification of TBEV-EU, formerly called central European encephalitis autochthonous cases. virus (5). Five of 6 ELISA positive samples and 1 of 7 borderline samples were confirmed positive by SNT. Five of ick-borne encephalitis virus (TBEV) can infect hu- the 6 SNT-confirmed roe deer were shot at or near a popu- Tmans, causing febrile illness; neurologic complications lar recreation area, the National Park Sallandse Heuvelrug include encephalitis (1). TBEV is transmitted through bites (Figure, panel A). of infected ticks to many animals, including deer, which In response to the serologic findings, we collected serve as feeding hosts for ticks (2,3). Expansion of TBEV 1,160 nymph and 300 adult Ixodes ricinus ticks by blan- subtypes has been reported (4). Reports of TBEV-neutral- ket dragging in 7 locations at the national park in Septem- izing antibodies in wildlife and cattle in Belgium prompted ber 2015. We extracted RNA from pools of 5 nymphs or 2 us to reinvestigate the presence of TBEV in the Nether- adults (7) and tested for flavivirus by using a reverse tran- lands (5,6). scription quantitative PCR. We detected flavivirus RNA in During January–September 2010, hunters collected 1 nymph pool and 1 pool of adult female ticks. 297 blood samples from roe deer (Capreolus capreolus) To obtain sequences of the 2 reverse transcription from locations across the Netherlands. We used a com- quantitative PCR–positive samples, we used primers and mercial ELISA to detect TBEV-reactive antibodies in roe protocols as described (8). Both sequences obtained from

Figure. Spatial distribution of TBEV-positive roe deer and genetic cluster analysis of TBEV sequences from the Netherlands. A) Spatial distribution of serologic test results (solid black circle, SNT positive; open white circle, ELISA and/or SNT negative) for 297 serum samples from roe deer collected according to a sampling scheme designed to obtain a representative sample of the roe deer population from locations across the Netherlands. Enlargement of the National Park Sallandse Heuvelrug area indicates the locations of the TBEV serologically positive roe deer (solid black circle) in relation to the site with reverse transcription quantitative PCR–positive ticks (solid black star) from 2015. B) Genetic cluster analysis of TBEV-NL sequences obtained from tick pools in the Netherlands with other tickborne viruses (indicated by GenBank accession number). Bold indicates the TBEV-NL sequence, which consists of 10,242 nt of the genome (GenBank accession no. LC171402), and the TBEV-EU strain with which TBEV-NL clusters. Where available, representatives of the subtypes are included. We conducted distance-based analyses using Kimura 2-parameters distance estimates and constructed the trees using the neighbor-joining algorithm, implemented in Bionumerics 7.1 (Applied Math, Sint-Martens-Latem, Belgium). We calculated bootstrap proportions by analyzing 1,000 replicates for neighbor-joining trees. Scale bar indicates nucleotide substitutions per site. GGEV, Greek goat encephalomyelitis virus; LIV, Louping ill virus; SGEV, Spanish goat encephalitis virus; SSEV, Spanish sheep encephalitis virus; TBEV, tickborne encephalitis virus; TBEV-EU, TBEV European subtype; TBEV-FE, Far Eastern subtype; TBEV-NL, TBEV Netherlands subtype; TBEV-SI, TBEV Siberian subtype; TSEV, Turkish sheep encephalomyelitis virus.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1029 RESEARCH LETTERS the tick pools were identical. The sequences obtained in Ms. Jahfari is a PhD candidate at the Dutch National Institute for this study were designated TBEV-NL and clustered within Public Health and the Environment (RIVM) and Erasmus Medical the TBEV-EU subtype complex (Figure, panel B), with a Center. Her primary research interest is tickborne diseases. 91% sequence identity with the currently known TBEV- EU sequences. References TBEV-EU RNA in 2 pools of ticks collected through 1. Lindquist L, Vapalahti O. Tick-borne encephalitis. Lancet. surveillance in 1 national park confirms the presence of 2008; 371:1861–71. http://dx.doi.org/10.1016/S0140-6736 TBEV-EU in the Netherlands. Serologic evidence that roe (08)60800-4 2. Bakhvalova VN, Dobrotvorsky AK, Panov VV, Matveeva VA, deer from the same location had been infected with a flavi- Tkachev SE, Morozova OV. Natural tick-borne encephalitis virus virus, most probably a TBEV, 5 years before the detection infection among wild small mammals in the southeastern part of of TBEV RNA in ticks suggests that TBEV has been en- western Siberia, Russia. Vector Borne Zoonotic Dis. 2006;6:32–41. demic to the Netherlands for at least 5 years. http://dx.doi.org/10.1089/vbz.2006.6.32 3. Gerth HJ, Grimshandl D, Stage B, Döller G, Kunz C. Roe deer The concentration of serologically positive roe deer is as sentinels for endemicity of tick-borne encephalitis virus. striking and remains unexplained. One explanation could Epidemiol Infect. 1995;115:355–65. http://dx.doi.org/10.1017/ be that this area has dense beech tree coverage, and beech- S0950268800058477 nuts are a major food source for roe deer and the bank vole 4. Donoso Mantke O, Schädler R, Niedrig M. A survey on cases of tick-borne encephalitis in European countries. Euro Surveill. 2008; (Myodes glareolus). These host species play a pivotal role 13:18848. in the TBEV enzootic cycle; a habitat suitable for both may 5. Roelandt S, Suin V, Van der Stede Y, Lamoral S, Marche S, have enhanced the local establishment and spread of TBEV. Tignon M, et al. First TBEV serological screening in Flemish wild In addition, the finding of a serologically positive roe deer boar. Infect Ecol Epidemiol. 2016;6:31099. http://dx.doi.org/ 10.3402/iee.v6.31099 in a southern province of the Netherlands (Figure, panel A), 6. Roelandt S, Suin V, Riocreux F, Lamoral S, Van der Heyden S, also known for the presence of beech trees, suggests that Van der Stede Y, et al. Autochthonous tick-borne encephalitis TBEV is distributed more widely within the Netherlands. virus–seropositive cattle in Belgium: a risk-based targeted Dissemination of information about the occurrence serological survey. Vector Borne Zoonotic Dis. 2014;14:640–7. http://dx.doi.org/10.1089/vbz.2014.1576 of TBEV in ticks and wildlife is needed for medical 7. Klaus C, Hoffmann B, Hering U, Mielke B, Sachse K, Beer M, professionals and the general public. In response to our et al. Tick-borne encephalitis (TBE) virus prevalence and findings, 2 autochthonous TBEV infections were report- virus genome characterization in field-collected ticks Ixodes( ricinus) ed in the Netherlands (9,10). At least 1 of these autoch- from risk, non-risk and former risk areas of TBE, and in ticks removed from humans in Germany. Clin Microbiol thonous cases was infected with a TBEV strain show- Infect. 2010;16:238–44. http://dx.doi.org/10.1111/ ing 99% homology with the Neudörfl strain, suggesting j.1469-0691.2009.02764.x the presence of multiple TBEV-EU strains in the Neth- 8. Kupča AM, Essbauer S, Zoeller G, de Mendonça PG, erlands. Our findings indicate that clinicians should be Brey R, Rinder M, et al. Isolation and molecular characterization of a tick-borne encephalitis virus strain from a new tick-borne aware of the possibility for TBEV infection in humans encephalitis focus with severe cases in Bavaria, Germany. in the Netherlands. Ticks Tick Borne Dis. 2010;1:44–51. http://dx.doi.org/10.1016/ j.ttbdis.2009.11.002 9. de Graaf JA, Reimerink JHJ, Voorn GP, bij de Vaate EA, Acknowledgments de Vries A, Rockx B, et al. First human case of tick-borne We thank Fedor Gassner, Gilian van Duijvendijk , Ryanne encephalitis virus infection acquired in the Netherlands, July 2016 Jaarsma, Aleksandra Krawczyk, and Miriam Maas for performing Euro Surveill. 2016;21:30318. http://dx.doi.org/10.2807/1560- fieldwork; Daan Vreugdenhil and Tom Klomphaar for access to 7917.ES.2016.21.33.30318 10. Weststrate AC, Knapen D, Laverman GD, Schot B, Prick JJ, the nature reserves; Natashja Buijs, Ewa Frazer, Najima Lamkaraf, Spit SA, et al. Increasing evidence of tick-borne encephalitis and Sophie Lamoral for technical support in the laboratory; and (TBE) virus transmission, the Netherlands, June 2016. Euro Marion Koopmans for critically reading this manuscript. Surveill. 2017;22:30482. http://dx.doi.org/10.2807/1560-7917. ES.2017.22.11.30482 This study was supported by the Netherlands Ministry of Health, Welfare, and Sport and performed under the frame of EurNegVec Address for correspondence: Hein Sprong, Center for Infectious Disease Cost Action TD1303. The collection of roe deer sera in 2010 Control, National Center for Public Health and the Environment, Antonie was financed by the Netherlands Ministry of Economic Affairs van Leeuwenhoeklaan 9, 3721 MA, Bilthoven, the Netherlands; email: (former LNV; verplichtingnummer 1400004212). [email protected]

1030 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS

Outbreaks of Tilapia Lake northeastern Thailand (online Technical Appendix Fig- ure 1, https://wwwnc.cdc.gov/EID/article/23/6/16-1278- Virus Infection, Thailand, Techapp1.pdf). Affected fish were commonly found within 2015–2016 1 month after transfer from the hatchery facility to grow- out ponds or cages. In general, clinical signs and high mortality rates were associated with fish weighing 1–50 g (on- Win Surachetpong, Taveesak Janetanakit, line Technical Appendix Figure 2). Mortality rates among Nutthawan Nonthabenjawan, tilapia farms were 20%–90%; higher rates were associat- Puntanat Tattiyapong, Kwanrawee Sirikanchana, ed with secondary bacterial and parasitic infections. Mor- Alongkorn Amonsin tality rates peaked within 14 days after the first dead fish Author affiliations:, Kasetsart University, Bangkok, Thailand were found. (W. Surachetpong, P. Tattiyapong); Chulalongkorn University, As part of the outbreak investigation, samples of Bangkok (T. Janetanakit, N. Nonthabenjawan, A. Amonsin); brain tissue were taken from fish at each of the 32 out- Chulabhorn Research Institute, Bangkok (K. Sirikanchana); break locations (each with a mortality rate >1%/day for Ministry of Education, Bangkok (K. Sirikanchana) 3 consecutive days): 10–30 moribund fish and 5–10 ap- parently healthy fish from the same culture areas. In to- DOI: https://dx.doi.org/10.3201/eid2306.161278 tal, 325 samples were collected and tested for etiologic agent(s) (4,6). Samples from fish involved in 22 of the 32 During 2015–2016, several outbreaks of tilapia lake virus outbreaks were positive for TiLV (online Technical Ap- infection occurred among tilapia in Thailand. Phylogenet- pendix Table 1). ic analysis showed that the virus from Thailand grouped For our study, we selected a field sample positive for with a tilapia virus (family Orthomyxoviridae) from Israel. TiLV (designated TiLV/Tilapia/Thai/TV1/2016) and pro- This emerging virus is a threat to tilapia aquaculture in Asia cessed it for whole-genome sequencing. Another 6 TiLVs and worldwide. were selected for sequencing of the putative polymerase basic 1 (PB1) gene (online Technical Appendix Table 2). iral diseases are common causes of illness and TiLV genome sequencing was conducted by using newly Vdeath in cultured fish; such viruses include infec- designed primers based on reference TiLVs available in the tious salmon anemia virus, infectious hematopoietic ne- GenBank database (7). Nucleotide sequences of 7 TiLVs crosis virus, and viral hemorrhagic septicemia virus (1). from Thailand were submitted to GenBank (accession nos. With regard to tilapia, some viral pathogens, including KX631921–36). betanodavirus, iridovirus, and herpes-like virus (2,3), re- Comparison of the TiLVs from Thailand with those portedly cause severe disease. In recent years, Thailand from Israel showed high nucleotide and amino acid iden- has experienced extensive losses of tilapia; most losses tities (95.18%–99.10%). Among TiLVs from Thailand, occurred 1 month after transfer of fish from hatchery to nucleotide and amino acid identities for segment 1 or the grow-out cages in public rivers or reservoirs (1-month putative PB1 gene of the virus were high (99.61%–100%) mortality syndrome). During routine investigation of (online Technical Appendix Table 3). Genetic analysis of this syndrome, multiple bacterial and parasitic infections the putative PB1 protein of TiLVs from Thailand and the were identified. However, no association was -estab viruses of the family Orthomyxoviridae showed that TiL- lished between the outbreaks and any primary causative Vs from Thailand possessed motifs preA, A, B, C, D, and agent(s). Most deaths occurred within 2 weeks after the E similar to those of Orthomyxoviridae viruses, including first dead fish were found. Similar observations of -ex influenza A, B, and C viruses; infectious salmon anemia tensive losses of raised tilapia and wild fish in Israel and virus; Dhori virus; and Thogoto virus (online Technical Ecuador have been reported (4,5). These outbreaks led Appendix Table 4) (8–10). Phylogenetic analysis showed to identification of a virus affecting tilapia, called tilapia that TiLVs from Thailand were closely related to TiLVs lake virus (TiLV). The epidemiologic pattern and clini- from Israel and grouped with the viruses of the family cal signs for infected fish in Thailand led to suspicion Orthomyxoviridae but not Arenaviridae and Bunyaviridae of an illness of unknown etiology that was similar to (Figure). This result suggests that the genetic composi- TiLV infection. tion of this emerging virus was similar to that of ortho- During 2015–2016, we investigated 32 out- myxoviruses and homologous with previously published breaks involving a large number of deaths of unknown TiLV sequences. cause among Nile tilapia (Oreochromis niloticus) and Our PCR and whole-genome findings demonstrate red hybrid tilapia (Oreochromis spp.). The outbreaks genetic homology between TiLV from Thailand and occurred at fish farms in central, western, eastern, and the etiologic agent of a novel RNA virus infection of

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1031 RESEARCH LETTERS

Figure. Phylogenetic analysis of the nucleotide sequences of RNA polymerase of TiLVs from Thailand (triangles) and reference viruses of the families Orthomyxoviridae, Arenaviridae, and Bunyaviridae. Genus and family groups are indicated; GenBank accession numbers are provided for reference viruses. The phylogenetic tree was constructed by using MEGA 6.0 (10) and applying a neighbor-joining bootstrap analysis (1,000 replications) with the Poisson model and gamma distribution. Human respiratory syncytial virus was used as an outgroup. ISAV, infectious salmon anemia virus; PUUV, Puumala virus; TiLV, tilapia lake virus. Scale bar indicates nucleotide substitutions per site. tilapia in Israel and Ecuador (4,7). Furthermore, the clin- proteins, including segment 1 or putative PB1 protein. ical signs and pathological presentation of infection with The pattern of protein motifs for this putative PB1 was TiLV from Thailand are similar to those of infection with similar to that for influenza viruses. To our knowledge, TiLV from Israel (online Technical Appendix Figure 2). TiLV has infected tilapia only, no other aquatic or ter- The clinical signs, gross lesions, and histopathologic le- restrial animals. sions combined with virus identification and character- Our results emphasize that the virus isolated from ization highlight emerging TiLV in Thailand as the pri- Thailand shares high sequence similarity with TiLV from mary cause of the outbreaks. We also found that fish that Israel, suggesting that this virus spreads across continents. survived massive die-offs rarely showed clinical signs, Given that tilapia are the main aquaculture species, control suggesting the development of specific immunity of TiLV will be improved by further efforts such as strict against the virus. It should be noted that the TiLVs biosecurity, vaccine development, and selection of resistant from Thailand possessed 10 gene segments encoding 10 tilapia breeds.

1032 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS

Acknowledgments Endemic Hantavirus in Field We thank the National Research Council of Thailand and the Agricultural Research Development Agency for financial support Voles, Northern England (PRP6005020450) and the Center for Advanced Studies for Agriculture and Food, Kasetsart University Institute for Advanced Anna G. Thomason, Michael Begon, Studies. We also thank Chulalongkorn University for financial Janette E. Bradley, Steve Paterson, support to the Center of Excellence for Emerging and Joseph A. Jackson Re-emerging Infectious Diseases in Animals, the Thailand Research Fund for financial support to the Thailand Research DOI: https://dx.doi.org/10.3201/eid2306.161607 Fund Senior Scholar (to A.A.) (RTA5780006), and the National Science and Technology Development Agency (P-15-50004). Author affiliations: University of Salford, Salford, UK (A.G. Thomason, J.A. Jackson); University of Liverpool, Liverpool, Dr. Surachetpong is an assistant professor at the Faculty of UK (M. Begon, S. Paterson); University of Nottingham, Veterinary Medicine, Kasetsart University, Bangkok. His Nottingham, UK (J.E. Bradley) research interests are emerging viruses and immunology in We report a PCR survey of hantavirus infection in an ex- aquatic animals. tensive field vole (Microtus agrestis) population present in the Kielder Forest, northern England. A Tatenale virus–like References lineage was frequently detected (≈17% prevalence) in liver 1. Crane M, Hyatt A. Viruses of fish: an overview of significant tissue. Lineages genetically similar to Tatenale virus are pathogens. Viruses. 2011;3:2025–46. http://dx.doi.org/10.3390/ likely to be endemic in northern England. v3112025 2. Keawcharoen J, Techangamsuwan S, Ponpornpisit A, Lombardini ED, Patchimasiri T, Pirarat N. Genetic characterization ecently a new vole-associated hantavirus (Tatenale vi- of a betanodavirus isolated from a clinical disease outbreak in Rrus) was discovered in northern England (1), but only farm-raised tilapia Oreochromis niloticus (L.) in Thailand. J Fish from an individual Microtus agrestis field vole. Previously Dis. 2015;38:49–54. http://dx.doi.org/10.1111/jfd.12200 only hantaviruses from murine-associated lineages (Seoul 3. Shlapobersky M, Sinyakov MS, Katzenellenbogen M, Sarid R, Don J, Avtalion RR. Viral encephalitis of tilapia larvae: primary virus [SEOV] and SEOV-like viruses) had been reported characterization of a novel herpes-like virus. Virology. in the United Kingdom, despite the abundance of potential 2010;399:239–47. http://dx.doi.org/10.1016/j.virol.2010.01.001 vole hosts in the mainland United Kingdom and the en- 4. Eyngor M, Zamostiano R, Kembou Tsofack JE, Berkowitz A, demicity of vole-associated hantavirus lineages (Puumala Bercovier H, Tinman S, et al. Identification of a novel RNA virus lethal to tilapia. J Clin Microbiol. 2014;52:4137–46. virus [PUUV] and Tula virus) in mainland Europe (2). Here http://dx.doi.org/10.1128/JCM.00827-14 we present data suggesting that the Tatenale virus lineage is 5. Ferguson HW, Kabuusu R, Beltran S, Reyes E, Lince JA, endemic in northern England. del Pozo J. Syncytial hepatitis of farmed tilapia, Oreochromis European hantaviruses are of public health signifi- niloticus (L.): a case report. J Fish Dis. 2014;37:583–9. http://dx.doi.org/10.1111/jfd.12142 cance because they are a causative agent of hemorrhagic fe- 6. Kurita J, Nakajima K, Hirono I, Aoki T. Polymerase chain ver with renal syndrome (HFRS). In the United Kingdom, reaction (PCR) amplification of DNA of red seam bream iridovirus HFRS cases have primarily been attributed to SEOV-like (RSIV). Fish Pathology. 1998;33:17–23. http://dx.doi.org/10.3147/ viruses on the basis of serologic tests. SEOV antibodies jsfp.33.17 7. Bacharach E, Mishra N, Briese T, Zody MC, Kembou Tsofack JE, have been detected in both humans and Norway rats (Rat- Zamostiano R, et al. Characterization of a novel orthomyxo-like tus norvegicus) in Northern Ireland and Yorkshire (3,4), virus causing mass die-offs of tilapia. MBio. 2016;7:e00431–16. and seropositivity in humans correlates with domestic or http://dx.doi.org/10.1128/mBio.00431-16 occupational exposure to rats (3,5). However, in the United 8. Müller R, Poch O, Delarue M, Bishop DH, Bouloy M. Rift Valley fever virus L segment: correction of the sequence and possible Kingdom, HFRS cases with serologic cross-reactivity to functional role of newly identified regions conserved in RNA- PUUV (3), which might share antigenic determinants with dependent polymerases. J Gen Virol. 1994;75:1345–52. Tatenale virus, have occurred. http://dx.doi.org/10.1099/0022-1317-75-6-1345 To investigate the endemicity of hantavirus in field 9. Poch O, Sauvaget I, Delarue M, Tordo N. Identification of four conserved motifs among the RNA-dependent polymerase encoding voles in the United Kingdom, we surveyed the extensive elements. EMBO J. 1989;8:3867–74. field vole population in the Kielder Forest, Northumberland 10. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: (≈230 km distant from the locality where Tatenale virus was Molecular Evolutionary Genetics Analysis version 6.0. Mol Biol discovered). All sampled sites were grassy, clear-cut areas Evol. 2013;30:2725–9. http://dx.doi.org/10.1093/molbev/mst197 (adjacent to forest stands) where field voles were prevalent. Address for correspondence: Win Surachetpong, Faculty of Veterinary Fieldwork was approved by the University of Liverpool Medicine, Kasetsart University, Bangkok, Thailand 10900; email: Animal Welfare Committee and conducted subject to UK [email protected] home office project license PPL 70_8210. Following the

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1033 RESEARCH LETTERS capture and processing of animals as previously described independent replicate PCRs) by Sanger sequencing (Source (6), we extracted viral RNA from 48 livers using a QIAamp BioScience, Rochdale, UK). The 380-bp sequence was Viral RNA Mini Kit (QIAGEN, Manchester, UK) and con- identical in all 3 positive vole samples, with the exception verted to cDNA using a High-Capacity RNA-to-cDNA of a single nucleotide polymorphism at position 145 (ad- Kit (Applied Biosystems, Warrington, UK). Hantaviruses enine, 2 voles; guanine, 1 vole; GenBank accession nos. were detected by PCR amplification of a fragment of the KY751731, KY751732). The recovered sequences were genomic L segment encoding RNA polymerase, following similar to but divergent from Tatenale virus (86.0%–86.3% the strategy outlined by Klempa et al. (7). identity at the nucleotide level and 95.9%–96.7% identity PCR-positive results were recorded for 16.7% of voles at the amino acid level; online Technical Appendix Figures (8/48) in total; positive voles were recorded at 3 of the 5 1, 2). Phylogenetic analysis of the L segment demonstrated sampled sites and throughout the survey period, March– this level of divergence was comparable to the divergence September 2015 (Figure, panel A; online Technical Ap- among many western European lineages of PUUV (Figure, pendix Table, https://wwwnc.cdc.gov/EID/article/23/6/16- panel B; online Technical Appendix Figure 3). 1607-Techapp1.pdf). Three positive samples from different Taken together with the original report, these data individual voles were sequenced (in both directions from are sufficient to suggest that Tatenale-like hantavirus

Figure. Investigation of Tatenale virus–like hantavirus lineages in the United Kingdom. A) Locations on the UK mainland where Tatenale virus–like hantavirus lineages have been found: Tattenhall, Cheshire, the site where Tatenale virus was discovered (1); and Kielder Forest, Northumberland. Kielder Forest sample sites are indicated in the inset (Geordies Knowe [GRD]: 55°11′1.61′′N, 2°35′3.05′′W; Cheese Sike [CHE]: 55°13′8.39′′N, 2°32′26.50′′W; Scaup [SCP]: 55°15′44.18′′N, 2°32′41.05′′W). B) Bayesian phylogenetic tree for the hantavirus genomic large segment (318-bp fragment with no missing data), showing relationships among Tatenale virus-like lineages and other relevant lineages. Bold represents the Tatenale virus–like lineage found in this study; either sequence reported here produces the same tree. Phylogenetic analysis was conducted by using a general time reversible plus gamma plus invariant sites model within MrBayes (8) software using Markov chain Monte Carlo chain lengths of 1 million and a strict clock. We estimated substitution models using MrModelTestV2 (9). Sequences are represented by the taxonomic names, strain (if >1 is included), and GenBank accession numbers. The tree is drawn to scale with node values representing the posterior probabilities. Scale bar represents nucleotide substitutions per site. ADLV, Adler virus; CATV, Catacamas virus; HOKV, Hokkaido virus; KBRV, Khabarovsk virus; LNV, Laguna Negra virus; PHV, Prospect Hill virus; PUUV, Puumala virus; TULV, Tula virus.

1034 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS lineages are widespread and common in northern Eng- 8. Ronquist F, Huelsenbeck JP. MrBayes 3: Bayesian phylogenetic land. Furthermore, the considerable sequence diver- inference under mixed models. Bioinformatics. 2003;19:1572–4. http://dx.doi.org/10.1093/bioinformatics/btg180 gence between samples in Cheshire and Northumberland 9. Nylander J. MrModeltest v2. Uppsala: Evolutionary Biology is consistent with a long-standing endemicity in northern Centre, Uppsala University; 2004. England. Given that PUUV has never been recorded in 10. Pether JV, Lloyd G. The clinical spectrum of human hantavirus the United Kingdom (2,10), the possibility should be infection in Somerset, UK. Epidemiol Infect. 1993;111:171–5. http://dx.doi.org/10.1017/S095026880005679X considered that a Tatenale-like virus could have been responsible for some of the HFRS cases that have oc- Address for correspondence: Joseph A. Jackson, University of Salford, curred here. More studies are needed to confirm whether School of Environment and Life Sciences, Peel Building, Salford M5 other common rodents in the United Kingdom are hosts 4WT, UK; email: [email protected] for this virus and to further characterize its phyletic relationships and zoonotic potential. Cross-reactivity of the sera from Tatenale-like virus–infected individuals to antigens of other relevant hantaviruses should be determined to inform future serologic surveys and the diagnosis of human HFRS cases.

Acknowledgments Measles Cases during Ebola We thank Rebecca Barber, Lukasz Lukomski, Steve Price, and Outbreak, West Africa, Ann Lowe for their assistance. 2013–2106 This work was funded by the Natural Environment Research Council, United Kingdom (research grant NE/L013517/2). Francesca Colavita,1 Mirella Biava,1 Ms. Thomason is a doctoral student at the School of Concetta Castilletti, Serena Quartu, Environment and Life Sciences at the University of Salford, Francesco Vairo, Claudia Caglioti, Chiara Agrati, Manchester, United Kingdom. Her interests are in the ecological Eleonora Lalle, Licia Bordi, Simone Lanini, dynamics of infectious disease in wildlife. Michela Delli Guanti, Rossella Miccio, Giuseppe Ippolito, Maria R. Capobianchi, Antonino Di Caro; and the Lazzaro Spallanzani References 1. Pounder KC, Begon M, Sironen T, Henttonen H, Watts PC, Institute for Research and Health Care Ebola 2 Voutilainen L, et al. Novel hantavirus in wildlife, United Virus Disease Sierra Leone Study Group Kingdom. Emerg Infect Dis. 2013;19:673–5. http://dx.doi.org/ Author affiliations: Lazzaro Spallanzani Institute for Research 10.3201/eid1904.121057 2. Bennett E, Clement J, Sansom P, Hall I, Leach S, Medlock JM. and Health Care, Rome, Italy (F. Colavita, M. Biava, C. Castilletti, Environmental and ecological potential for enzootic cycles S. Quartu, F. Vairo, C. Caglioti, C. Agrati, E. Lalle, L. Bordi, of Puumala hantavirus in Great Britain. Epidemiol Infect. S. Lanini, G. Ippolito, M.R. Capobianchi, A. Di Caro); Emergency 2010;138:91–8. http://dx.doi.org/10.1017/S095026880999029X Nongovernmental Organization, Milan, Italy (M. Delli Guanti, 3. Jameson LJ, Newton A, Coole L, Newman EN, Carroll MW, Beeching NJ, et al. Prevalence of antibodies against hantaviruses R. Miccio) in serum and saliva of adults living or working on farms in Yorkshire, United Kingdom. Viruses. 2014;6:524–34. DOI: https://dx.doi.org/10.3201/eid2306.161682 http://dx.doi.org/10.3390/v6020524 4. McKenna P, Clement J, Matthys P, Coyle PV, McCaughey C. The recent Ebola outbreak in West Africa caused breakdowns Serological evidence of hantavirus disease in Northern Ireland. in public health systems, which might have caused outbreaks J Med Virol. 1994;43:33–8. http://dx.doi.org/10.1002/ jmv.1890430107 of vaccine-preventable diseases. We tested 80 patients admit- 5. Jameson LJ, Taori SK, Atkinson B, Levick P, Featherstone CA, ted to an Ebola treatment center in Freetown, Sierra Leone, van der Burgt G, et al. Pet rats as a source of hantavirus in England for measles. These patients were negative for Ebola virus. and Wales, 2013. Euro Surveill. 2013;18:18. Measles virus IgM was detected in 13 (16%) of the patients. 6. Jackson JA, Begon M, Birtles R, Paterson S, Friberg IM, Hall A, et al. The analysis of immunological profiles in wild animals: a case study on immunodynamics in the field vole,Microtus he Ebola virus disease (EVD) outbreak in West Africa agrestis. Mol Ecol. 2011;20:893–909. http://dx.doi.org/10.1111/ during 2013–2016 was one of the worst public health di- j.1365-294X.2010.04907.x T 7. Klempa B, Fichet-Calvet E, Lecompte E, Auste B, Aniskin V, sasters in recent history; it caused >28,646 cases and 11,323 Meisel H, et al. Hantavirus in African wood mouse, Guinea. Emerg Infect Dis. 2006;12:838–40. http://dx.doi.org/10.3201/ 1These authors contributed equally to this article. eid1205.051487 2Members of this group are listed at the end of this article.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1035 RESEARCH LETTERS deaths (1). Consequences of EVD include social instability, We analyzed 80 patients, of whom 27 were >8 years poor food reserves, breakdown of healthcare systems, and of age and <25 years of age (median age 30 years) during reduced vaccination coverage (2,3). Breakdown of health- December 2014–June 2015, who had fever (temperature care systems and reduced vaccination coverage might have >37.5°C) and diarrhea or vomiting. Only 1 patient had a his- been the worst consequences because nearly all health re- tory of rash. Measles virus IgM was detected by using an in- sources were shifted to the EVD response. Disruptions of direct immunofluorescence assay for 13 patients (16%), most local health systems could lead to underreporting of other (69%) of whom were in this age group (p<0.001). Although infectious diseases cases and a second crisis that could kill as we could not determine whether measles IgM in these pa- many as persons as the original outbreak, if not more. tients was caused by vaccination failures during the outbreak The 3 countries most affected by this outbreak (Sierra or failures of the Expanded Programme on Immunization, Leone, Guinea, and Liberia) have been a major part of the this age distribution might be indicative of noneffective im- World Health Organization Expanded Programme on Im- munization campaigns, especially a deficiency in the health- munization through vaccination campaigns for reducing care system in detecting and reporting measles cases. childhood deaths from vaccine-preventable diseases, such as The recent EVD outbreak caused breakdowns of measles. Although there are no cures for EVD or measles, healthcare systems in the affected countries, leading to pos- a potent measles vaccine is available, which can prevent sible secondary outbreaks (2,3). A higher risk for vaccine- spread of this disease. Use of this vaccine is crucial because preventable diseases, in particular measles, is often an early measles is far more contagious (1 case-patient with measles result in interruption in delivery of public health services. can transmit infection to 12–18 persons) than EVD and Recent studies have shown that the increase in measles cas- might be the primary cause of major epidemics (3,4). These es during the EVD outbreak in 2013–2016 was caused by 3 countries reported nearly 93,685 cases of measles during disruption of vaccination programs and underreporting of 1994–2003 (although Sierra Leone did not report cases for 4 measles cases, which is probably related to effects of EVD years), and during November 1, 2009–July 13, 2010, a total on healthcare systems (9). of 1,094 confirmed measles cases were reported in Sierra Le- Our results for samples obtained from Ebola-negative one. Plans for measles vaccination campaigns were imple- patients showed a high number of measles infections dur- mented before the EVD outbreak because of an increase in ing the outbreak in different age groups. Although few (n susceptibility to measles in these 3 countries (3,5). = 80) patients have been tested, our results provide useful Historically, measles outbreaks have followed hu- insights into measles cases during other outbreaks in dif- manitarian crises, such as war (6), natural disasters (7), and ferent age groups, adding new evidence from a study that political crises (8). Recent studies have shown that measles focused on children (9). is one of the causative agents of secondary outbreaks dur- Our findings indicate the need for correct and rapid dif- ing the EVD epidemics in West Africa (9), probably due ferential diagnoses during such outbreaks to avoid spread to the disruption in vaccination campaigns, nonfunctional of other infectious diseases. Furthermore, local public healthcare systems (including detection and reporting of health systems should be strengthened in those countries measles cases), lack of specific treatment, and a sense of that are now recovering from the EVD outbreak to reduce fear of contracting EVD with reluctance to approach health risks for other infectious diseases outbreaks. assistance (10). Members of the Lazzaro Spallanzani Institute for Research and Probable underreporting of and lack of data for mea- Health Care Ebola Virus Disease Sierra Leone Study Group: sles cases during EVD outbreaks prompted us to investigate Antonella Vulcano, Francesca Colavita, Carolina Venditti, measles in Sierra Leone during the recent EVD outbreak. Paola Zaccaro, Antonio Mazzarelli, Concetta Castilletti, Angela Although during the preparedness phase of the European Cannas, Serena Quartu, Sabrina Coen, Silvia Meschi, Claudia Mobile Laboratory Project (http://www.emlab.eu), measles Minosse, Roberta Chiappini, Mirella Biava, Maria Beatrice Valli, was to be included among diseases tested for differential Germana Grassi, and Daniele Lapa. diagnosis of EVD, we could not implement this approach in the field during the outbreak. We performed a retrospective serologic study to par- Acknowledgments tially investigate the role of measles in the EVD outbreak We thank Italian Ministry of Health (Ricerca Corrente and by testing serum samples negative for Ebola virus by re- Ricerca Finalizzata) for supporting deployment of laboratory verse transcription PCR for measles virus IgM from per- personnel; the Sierra Leone Ministry of Health and Sanitation, sons suspected of having EVD. Samples were obtained the Pharmacy Board of Sierra Leone, and the Sierra Leone at the Emergency Nongovernmental Organization Ebola National Laboratory Service for collaborating in overall Treatment Center (Goderich, Freetown, Sierra Leone). This laboratory activities implemented in the laboratory at the study was approved by Ethical Committee of Sierra Leone. Princess Christian Maternity Hospital (Freetown, Sierra

1036 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS

Leone); staff (nurses, physicians, pharmacists, logisticians, Address for correspondence: Antonino Di Caro, Laboratory of Virology, administrators, cleaners, drivers, cooks, guards) at the Lazzaro Spallanzani Institute for Research and Health Care, Via Emergency Ebola Treatment Center (Goderich, Freetown, Portuense 292, 00149 Rome, Italy; email: [email protected] Sierra Leone) and the Emergency Surgical and Pediatric Hospital (Goderich) for their efforts and support; and the people of Goderich for their energy, optimism, cordiality, and positive attitude.

This study was supported by the Italian Ministry of Foreign Affairs (Directorate General for Development Cooperation). Angiostrongylus cantonensis The Lazzaro Spallanzani Institute for Research and Health Care Meningitis and Myelitis, was supported by the European Union Horizon 2020 Research and Innovation Programme (EVIDENT, grant no. 666100); the Texas, USA European Union Innovative Medicine Initiative) Programme (projects EbolaMODRAD [grant no. 115843] and FILODIAG Roukaya Al Hammoud, Stacy L. Nayes, [grant no. 115844]); and the US Food and Drug Administration James R. Murphy, Gloria P. Heresi, Ian J. Butler, project (FDABAA-15-00121). Norma Pérez Author affiliation: McGovern Medical School, University of Texas Dr. Colavita is research scientist at the Laboratory of Virology, Health Science Center at Houston, Houston, Texas, USA Lazzaro Spallanzani Institute for Research and Health Care, Rome, Italy. Her research interests are infections with emerging DOI: https://dx.doi.org/10.3201/eid2306.161683 viruses and risk group 3 and 4 pathogens, host immune responses, and host–pathogen interactions. Infection with Angiostrongylus cantonensis roundworms is endemic in Southeast Asia and the Pacific Basin.A. cantonensis meningitis and myelitis occurred in summer 2013 in a child with References 1. World Health Organization. Ebola Situation report, 2015 no history of travel outside of Texas, USA. Angiostrongyliasis [cited 2017 Mar 29]. http://www.who.int/csr/disease/ebola/ is an emerging neurotropic helminthic disease in Texas and situation-reports/en/ warrants increased awareness among healthcare providers. 2. United Nations Development Group. Socio-economic impact of Ebola virus disease in West African countries, 2015. New York: United Nations Development Programme [cited 2017 Mar 29]. n summer 2013, a previously healthy Caucasian http://reliefweb.int/sites/reliefweb.int/files/resources/ebola- 12-month-old girl was brought for treatment to a chil- west-africa.pdf I 3. Takahashi S, Metcalf CJE, Ferrari MJ, Moss WJ, Truelove SA, dren’s hospital in Houston, Texas, USA, on the 11th day Tatem AJ, et al. Reduced vaccination and the risk of measles and of illness (day 11), manifesting intermittent fever, lethargy, other childhood infections post-Ebola. Science. 2015;347:1240–2. and emesis. She had been evaluated by a pediatrician on http://dx.doi.org/10.1126/science.aaa3438 day 3 and diagnosed with presumed viral infection. She at- 4. World Health Organization. Global vaccine action plan 2011–2020. 2013 [cited 2017 Mar 29]. http://www.who.int/immunization/ tended day care, had no history of sick contacts, and apart global_vaccine_action_plan/en/ from dogs in the house, had no notable other exposures. 5. Truelove SA, Moss WJ, Lessler J. Mitigating measles outbreaks At hospital admission, physical examination showed in West Africa post-Ebola. Expert Rev Anti Infect Ther. 2015; vital signs within reference ranges, mild distress, lethargy, 13:1299–301. http://dx.doi.org/10.1586/14787210.2015.1085305 6. Sharara SL, Kanj SS. War and infectious diseases: challenges and irritability with no focal deficits or signs of meningeal of the Syrian civil war. PLoS Pathog. 2014;10:e1004438. irritation. Blood test results showed leukocytosis (17,900 http://dx.doi.org/10.1371/journal.ppat.1004438 cells/mm3 with 20% eosinophils). Cerebrospinal fluid 7. Surmieda MR, Lopez JM, Abad-Viola G, Miranda ME, (CSF) examination showed 8 erythrocytes and 568 leuko- Abellanosa IP, Sadang RA, et al. Surveillance in evacuation camps 3 after the eruption of Mt. Pinatubo, Philippines. MMWR CDC cytes/mm with 26% eosinophils. Results of bacterial cul- Surveill Summ. 1992;41:9–12. tures and PCR of CSF for herpes simplex virus and entero- 8. UNICEF. Sanctions: children hard hit in Haiti [cited 2017 Mar 29]. virus were negative. She had no serologic evidence of acute http://www.unicef.org/sowc96/dsanctns.htm infection with West Nile virus or HIV. Magnetic resonance 9. Suk JE, Paez Jimenez A, Kourouma M, Derrough T, Baldé M, Honomou P, et al. Post-Ebola measles outbreak in Lola, imaging (MRI) of the brain showed normal results. She re- Guinea, January–June 2015. Emerg Infect Dis. 2016;22:1106–8. ceived ceftriaxone, vancomycin, and acyclovir from days http://dx.doi.org/10.3201/eid2206.151652 11 through 15 with no clinical improvement. 10. Shrivastava SR, Shrivastava PS, Jegadeesh R. Legacy of On day 16, because the child had been exposed to dogs, Ebola outbreak: potential risk of measles outbreak in Guinea, Sierra Leone and Liberia. J Res Med Sci. 2015;20:529–30. she was empirically treated for presumed Toxocara infec- http://dx.doi.org/10.4103/1735-1995.163982 tion with albendazole and prednisone for 5 days. Her clinical

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1037 RESEARCH LETTERS condition improved. However, 2 days after she completed reported human case of A. cantonensis meningitis in a non- the albendazole and prednisone regimen (day 23), fever and traveler in the contiguous United States occurred in 1995 in lethargy recurred, she was unable to bear weight, and CSF a 10-year-old boy in New Orleans, Louisiana, who had eaten analysis showed increased eosinophilia (1,500 leukocytes/ raw snails (4). Since the case we report, 2 additional cases in mm3 with 35% eosinophils). On day 23, MRI showed diffuse children were reported in Houston, in the spring of 2016 (5). leptomeningeal enhancement in the brain and spinal cord, The frequency of A. cantonensis infections in humans diffuse punctate areas of cortical infarction, and intramedul- is likely underreported due to a combination of factors, in- lary enhancement (Figure). On the same day, she was again cluding the frequent self-limited course of the infection, treated with albendazole and prednisone with good response. lack of awareness of the parasite, limited availability of Results of serologic testing performed at the hospital diagnostic testing, and lack of national surveillance. Al- were negative for Trichinella, Toxocara, and Schistosoma though the disease usually resolves spontaneously, case-fa- spp. Serum and CSF samples were sent to the Centers for tality rates can reach 5% (6). Lack of clinical suspicion for Diseases Control and Prevention (Atlanta, GA, USA); angiostrongyliasis on the basis of signs and symptoms and PCR results showed CSF was positive for Angiostrongylus delay in initiation of treatment can lead to neurologic dete- cantonensis, and serum was positive for Toxocara Stron- rioration, especially in young children, as demonstrated in gyloides stercoralis IgG. On day 31, we diagnosed A. can- our patient and another report (7). tonensis meningitis and myelitis with cross-reactions to In conclusion, angiostrongyliasis is an emerging neu- S. stercoralis and Toxocara spp. The patient completed 2 rotropic helminthic disease in Texas. Increased awareness weeks of albendazole therapy (from day 23 through day of A. cantonensis infection in humans is needed among 37) and 3 months of prednisone therapy with slow tapering. healthcare providers. The patient’s symptoms completely resolved by day 64. Dr. Al Hammoud is a pediatric infectious diseases fellow at Lumbar punctures on days 35, 56, and 104 showed gradual University of Texas Health, McGovern Medical School. Her return of CSF parameters to normal ranges. On day 56, re- areas of interest include general pediatric infectious diseases, sults of an MRI of the brain were unremarkable. HIV-associated nephropathy, and immunization response in The case was reported to the local health department, HIV-infected children. which tested snails trapped close to the patient’s residence. No evidence of A. cantonensis roundworms was found. References In 1987, A. cantonensis roundworm was identified for 1. Campbell BG, Little MD. The finding ofAngiostrongylus cantonensis the first time in North America in rodents. 1( ). Although a in rats in New Orleans. Am J Trop Med Hyg. 1988;38:568–73. 2. Teem JL, Qvarnstrom Y, Bishop HS, da Silva AJ, Carter J, PCR-based assay did not detect A. cantonensis roundworm White-McLean J, et al. The occurrence of the rat lungworm, in samples of intermediate hosts (snails) in Houston, Texas, Angiostrongylus cantonensis, in nonindigenous snails in the Gulf of the parasite has been repeatedly identified in southeastern Mexico region of the United States. Hawaii J Med Public Health. and Gulf states (Louisiana, Oklahoma, Mississippi, Florida) 2013;72(Suppl 2):11–4. 3. York EM, Creecy JP, Lord WD, Caire W. Geographic range (2,3). In parallel, human infections have increased. The first expansion for rat lungworm in North America. Emerg Infect Dis. 2015;21:1234–6. http://dx.doi.org/10.3201/eid2107.141980 4. New D, Little MD, Cross J. Angiostrongylus cantonensis infection from eating raw snails. N Engl J Med. 1995;332:1105–6. http://dx.doi.org/10.1056/NEJM199504203321619 5. Foster CE, Nicholson EG, Chun AC, Gharfeh M, Anvari S, Seeborg FO, et al. Angiostrongylus cantonensis infection: a cause of fever of unknown origin in pediatric patients. Clin Infect Dis. 2016;63:1475–8. http://dx.doi.org/10.1093/cid/ciw606 6. Tseng YT, Tsai HC, Sy CL, Lee SS, Wann SR, Wang YH, et al. Clinical manifestations of eosinophilic meningitis caused by Angiostrongylus cantonensis: 18 years’ experience in a medical center in southern Taiwan. J Microbiol Immunol Infect. 2011;44:382–9. http://dx.doi.org/10.1016/j.jmii.2011.01.034 7. Evans-Gilbert T, Lindo JF, Henry S, Brown P, Christie CD. Severe eosinophilic meningitis owing to Angiostrongylus cantonensis in young Jamaican children: case report and literature review. Paediatr Figure. Magnetic resonance imaging (MRI) of the brain (A) Int Child Health. 2014;34:148–52. http://dx.doi.org/10.1179/20469 and the spine (B) showing meningitis and myelitis in a 12-month- 05513Y.0000000106 old girl with Angiostrongylus cantonensis infection, Houston, Texas, USA. A) Axial T1 post contrast sequences showing diffuse Address for correspondence: Roukaya Al Hammoud, University of leptomeningeal enhancement (arrows). B) Sagittal T1 postcontrast Texas Health, McGovern Medical School, Pediatric Infectious Diseases sequences showing intramedullary enhancement in the thoracic and lumbar spinal cord T8–L5 with diffuse leptomeningeal Division, 6431 Fannin St, Ste 3.126, Houston, TX 77030, USA; enhancement (arrows). email: [email protected]

1038 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS

Enterocytozoon bieneusi profuse watery diarrhea and abdominal discomfort for 3 weeks. She had not traveled abroad. Three years earlier, Microsporidiosis in Stem Cell she received a genoidentical allogeneic HSCT for myeloid Transplant Recipients leukemia. She recently had cutaneous chronic graft-versus- Treated with Fumagillin1 host disease. Her immunosuppression regimen used was prednisone and mycophenolate mofetil. At admission, the patient was dehydrated and had a Iryna Bukreyeva, Adela Angoulvant, Inès Bendib, weight loss of 3 kg. Laboratory analyses showed lympho- Jean-Charles Gagnard, Jean-Henri Bourhis, cytopenia (960 lymphocytes/mm3), reference neutrophil Sylvie Dargère, Julie Bonhomme, Marc Thellier, (5,100 cells/mm3) and platelet (408,000 platelets/mm3) Bertrand Gachot, Benjamin Wyplosz counts, and a C-reactive protein level <5 mg/L. Author affiliations: Centre Hospitalier Universitaire de Bicêtre, Results of fecal sample analyses were negative for Le Kremlin-Bicêtre, France (I. Bukreyeva, A. Angoulvant, pathogenic bacteria and viruses. Microscopic examination I. Bendib, J.-C. Gagnard, B. Wyplosz); Université Paris-Sud, of fecal smears stained with Weber-Green–modified tri- Le Moulon, France (A. Angoulvant); Institut Gustave Roussy, chrome showed microsporidia. E. bieneusi was identified Villejuif, France (J.-H. Bourhis, B. Gachot); Centre Hospitalier by using monoclonal antibodies (IFA-MAbs; Bordier Af- Universitaire de Caen, Caen, France (S. Dargère, J. Bonhomme); finity Products, Crissier, Switzerland). Centre Hospitalier Universitaire Pitié-Salpêtrière, Paris, France The mycophenolate mofetil dose was reduced by (M. Thellier); Université Pierre et Marie Curie, Paris (M. Thellier) 50% for 8 days but no benefit was shown. Resolution of symptoms occurred <5 days after initiating fumagillin DOI: https://dx.doi.org/10.3201/eid2306.161825 therapy (60 mg/d for 14 d); fecal smears were negative for microsporidia on day 9, and transient thrombocytopenia Enterocytozoon bieneusi microsporidiosis is an emerg- (131,000 platelets/mm3) was observed on day 18 (Figure). ing disease in immunocompromised patients. We report Fecal smears remained negative for E. bieneusi during the 2 cases of this disease in allogeneic hematopoietic stem 6-month follow-up. No clinical relapse occurred. cell transplant patients successfully treated with fumagil- Patient 2 was a 42-year-old man referred to Centre lin. Thrombocytopenia occurred but without major adverse Hospitalier Universitaire de Bicêtre (Le Kremlin-Bicêtre, events. Modifications of immunosuppression could be France) after profuse acute diarrhea for 2 weeks and a avoided when E. bieneusi is rapidly identified and fumagillin therapy is started promptly. weight loss of 10 kg. He had not traveled abroad. Four years earlier, he received a genoidentical allogeneic HSCT for acute leukemia. During the follow-up period, he was nterocytozoon bieneusi, the most common cause of given a diagnosis of chronic graft-versus-host disease. He Emicrosporidiosis in humans (1), causes chronic diar- was given extracorporeal phototherapy with mycopheno- rhea and severe wasting syndrome in immunocompro- late mofetil, sirolimus, and prednisolone. mised patients (2). In 2002, oral fumagillin was established At admission, the patient was afebrile and dehydrat- as an effective treatment for E. bieneusi microsporidiosis ed. Blood analyses showed severe lymphocytopenia (400 in HIV-infected and solid organ transplant (SOT) patients lymphocytes/mm3), reference neutrophil (4,680 cells/ (3). In contrast to previous treatments that did not result mm3) and platelet (251,000 platelets/mm3) counts, and a in parasitologic clearance or clinical remission, fumagillin C-reactive protein level <5 mg/L. Results of microbiologi- showed a cure rate of 100%, even for severely immuno- cal analyses of fecal samples were negative for viruses compromised patients (2,4–10). and pathogenic bacteria. Microscopic examination of fe- Thrombocytopenia is the main adverse event of fuma- cal smears stained with Weber-Green–modified trichrome gillin therapy, occurring in up to 33% of patients (3) and showed microsporidia. E. bieneusi was identified by using raising concerns about fumagillin use in patients with he- monoclonal antibodies. matologic disorders. We report 2 cases of E. bieneusi mi- The patient was treated with fumagillin (60 mg/d for crosporidiosis in allogeneic hematopoietic stem cell trans- 14 d) (Figure). Immunosuppressive therapy was not modi- plant (HSCT) recipients who were treated with fumagillin fied. Clinical symptoms resolved within 5 days. Platelet and experienced thrombocytopenia. counts progressively decreased. Fumagillin was withdrawn Patient 1 was a 50-year-old woman admitted to Cen- on day 14, but thrombocytopenia worsened (40,000 plate- tre Hospitalier Universitaire de Caen (Caen, France) after lets/mm3) by day 18. However, the patient spontaneously 1Results from this study were presented at the 26th European recovered in 10 days without any bleeding. No relapses Congress of Clinical Microbiology and Infection Diseases; April were observed. Microsporidia were not detected in fecal 9–12, 2016; Amsterdam, the Netherlands. samples during the 6-month follow-up.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1039 RESEARCH LETTERS

Figure. Platelet counts and clinical and parasitologic characteristics during fumagillin therapy and a 1-month follow-up of 2 allogeneic hematopoietic stem cell recipients with Enterocytozoon bieneusi microsporidiosis.

E. bieneusi is an emerging pathogen in immunocom- patients we report, microsporidia were not detected in fecal promised patients (1). Increasing numbers of cases have samples of both patients during the 6-month follow-up. been reported in SOT patients. We report 2 cases of E. bi- In conclusion, fumagillin was highly efficient in cur- eneusi microsporidiosis in allogeneic HSCT recipients who ing E. bieneusi microsporidiosis in 2 allogeneic HSCT were treated with fumagillin without modifying the immu- recipients. Thrombocytopenia occurred but without major nosuppressive regimen for 1 recipient. In France, fumagillin adverse events. Modifications to immunosuppression could can be obtained from the French National Agency for Medi- be avoided when E. bieneusi is rapidly identified and fuma- cines and Health Products Safety (Saint-Denis, France) after gillin therapy is started promptly. an individual patient expanded-access request is submitted. Dr. Bukreyeva is a physician in the Infectious and Tropical Clinical and microbiological responses for the 2 case- Diseases Unit at the Bicêtre University Hospital, Kremlin- patients were similar to those reported for other immunocom- Bicêtre, France. Her research interests include prevention and promised patients (3). No relapses were observed for 4 HIV- treatment of infections in immunocompromised patients. infected patients whose CD4 cell counts remained low, or for 2 SOT recipients who did not receive tapering immunosuppressive therapy (3). In other studies, 15 (70%) of 21 patients References 1. Didier ES, Weiss LM. Microsporidiosis: current status. Curr treated with fumagillin were cured without modifying immu- Opin Infect Dis. 2006;19:485–92. http://dx.doi.org/10.1097/01. nosuppression regimens (2,6,10); for 6 other patients, immu- qco.0000244055.46382.23 nosuppressive therapy was tapered (n = 4) or withdrawn (n = 2. Lanternier F, Boutboul D, Menotti J, Chandesris MO, Sarfati C, 2), but reasons for modifying immunosuppression were not Mamzer Bruneel MF, et al. Microsporidiosis in solid organ transplant recipients: two Enterocytozoon bieneusi cases and specified. For 1 of our patients, the mycophenolate mofetil review. Transpl Infect Dis. 2009;11:83–8. http://dx.doi.org/10.1111/ dose was reduced by 50% to decrease the immunosuppres- j.1399-3062.2008.00347.x sion level. However, no benefit was observed. In contrast, 3. Molina JM, Tourneur M, Sarfati C, Chevret S, de Gouvello A, fumagillin led to clinical remission within 5 days. Gobert JG, et al.; Agence Nationale de Recherches sur le SIDA 090 Study Group. Fumagillin treatment of intestinal microsporidiosis. We observed thrombocytopenia (platelet count N Engl J Med. 2002;346:1963–9. http://dx.doi.org/10.1056/ 3 <40,000/mm ) in both patients but no evidence of bleeding. NEJMoa012924 In other non-AIDS patients, thrombocytopenia was report- 4. Desoubeaux G, Maakaroun-Vermesse Z, Lier C, Bailly E, ed in 11 (52%) SOT patients receiving fumagillin, includ- Morio F, Labarthe F, et al. Successful treatment with fumagillin of the first pediatric case of digestive microsporidiosis in a liver- ing 4 patients with severely low platelet counts (<25,000/ kidney transplant. Transpl Infect Dis. 2013;15:E250–9. 3 mm ) (1,4,6). For these patients, including those we re- http://dx.doi.org/10.1111/tid.12158 port, thrombocytopenia occurred during the second week 5. Pomares C, Santín M, Miegeville M, Espern A, Albano L, of treatment; a minimum value was observed a few days Marty P, et al. A new and highly divergent Enterocytozoon bieneusi genotype isolated from a renal transplant recipient. J Clin after completing fumagillin therapy. Spontaneous recovery Microbiol. 2012;50:2176–8. http://dx.doi.org/10.1128/JCM.06791-11 occurred within 2 weeks. Bleeding, hematoma, or require- 6. Godron A, Accoceberry I, Couret A, Llanas B, Harambat J. ments for platelet transfusions were not reported. For both Intestinal microsporidiosis due to Enterocytozoon bieneusi in

1040 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS

a pediatric kidney transplant recipient successfully treated with nfluenza A(H9N2) virus has been found in several avian fumagillin. Transplantation. 2013;96:e66–7. http://dx.doi.org/ species. Despite low pathogenicity in poultry, H9N2 vi- 10.1097/TP.0b013e3182a902e7 I 7. Kicia M, Wesolowska M, Jakuszko K, Kopacz Z, Sak B, ruses are important to public health because of their high Květonova D, et al. Concurrent infection of the urinary tract with adaptability and frequent infection in humans. Clinical cas- Encephalitozoon cuniculi and Enterocytozoon bieneusi in a renal es of H9N2 human infection have been reported in China transplant recipient. J Clin Microbiol. 2014;52:1780–2. (including Hong Kong) and Bangladesh (1,2). H9N2 viruss http://dx.doi.org/10.1128/JCM.03328-13 8. Ghoshal U, Khanduja S, Pant P, Prasad KN, Dhole TN, are constantly evolving and can reassort with other influ- Sharma RK, et al. Intestinal microsporidiosis in renal transplant enza A virus subtypes, resulting in novel influenza viruses. recipients: prevalence, predictors of occurrence and genetic H9N2 was the likely donor of internal genes for the H5N1, characterization. Indian J Med Microbiol. 2015;33:357–63. H7N9, and H10N8 viruses (3,4). http://dx.doi.org/10.4103/0255-0857.158551 9. Bednarska M, Bajer A, Welc-Faleciak R, Czubkowski P, During December 2014–August 2015, we conducted Teisseyre M, Graczyk TK, et al. The first case of Enterocytozoon an influenza A surveillance program in Shan State, Myan- bieneusi infection in Poland. Ann Agric Environ Med. 2013; mar. An outbreak of highly infectious avian influenza 20:287–8. A(H5N1) in November 2007 has been the only outbreak 10. Champion L, Durrbach A, Lang P, Delahousse M, Chauvet C, Sarfati C, et al. Fumagillin for treatment of intestinal reported in this state. For this study, we collected 648 sam- microsporidiosis in renal transplant recipients. Am J Transplant. ples from live-bird markets (LBMs) in Muse, Namkham, 2010;10:1925–30. http://dx.doi.org/10.1111/j.1600-6143.2010.03166.x Laukkai, and Chinshwehaw, Shan State townships on the China–Myanmar border (online Technical Appendix 1, Address for correspondence: Benjamin Wyplosz, Service de Maladies https://wwwnc.cdc.gov/EID/article/23/6/16-1902-Techapp1. Infectieuses et Tropicales, Centre Hospitalier Universitaire de Bicêtre, pdf). We collected oropharygeal swab specimens from 78 Rue du Général Leclerc, 94275 Le Kremlin-Bicêtre, France; chickens (n = 273) and ducks (n = 180) as well as environ- email: [email protected] mental samples (n = 195). Identification and isolation were performed at the Livestock Breeding and Veterinary Depart- ment, Yangon, Myanmar (online Technical Appendix). Of the 648 samples subjected to virus isolation by egg inoculation, 10 were hemagglutinin (HA) positive. We further confirmed 3 samples as influenza A virus by using real- time reverse transcription PCR (RT-PCR), and we subtyped Influenza A(H9N2) Virus, and confirmed all 3 as H9N2 (online Technical Appendix Myanmar, 2014–2015 Table 1). However, the overall occurrence of H9N2 in the LBMs in this study was relatively low. The 3 H9N2 isolates, A/chicken/Myanmar/NK-2/2015(H9N2), A/chicken/ Thant Nyi Lin, Nutthawan Nonthabenjawan, Myanmar/NK-4/2015(H9N2), and A/chicken/Myanmar/ Supassama Chaiyawong, Napawan Bunpapong, NK-5/2015(H9N2), were from chickens in LBMs in Nam- Supanat Boonyapisitsopa, Taveesak Janetanakit, kham Township in June 2015. Pont Pont Mon, Hla Hla Mon, Kyaw Naing Oo, To characterize the Myanmar H9N2 virus, we per- Sandi Myint Oo, Mar Mar Win, formed whole-genome sequencing of these 3 isolates and Alongkorn Amonsin submitted nucleotide sequences GenBank (accession nos. Author affiliations: Chulalongkorn University, Bangkok, Thailand KY115364–KY115387). Although an international stan- (T.N. Lin, N. Nonthabenjawan, S. Chaiyawong, N. Bunpapong, dard for clade nomenclature of the H9 subtype has not S. Boonyapisitsopa, T. Janetanakit, A. Amonsin); University of been well established, phylogenetic analysis showed that Veterinary Science, Nay Pyi Taw, Myanmar (T.N. Lin, S.M Oo, all 3 Myanmar H9N2 isolates were grouped into clade M.M. Win); Ministry of Agriculture, Livestock and Irrigation, Insein, 4.2.5 (HA gene) and BJ94-like (neuraminidase [NA] gene) Yangon, Myanmar (P.P. Mon); Ministry of Agriculture, Livestock (Figure; and online Technical Appendix Figure 2) (5). The and Irrigation, Sintgaing, Mandalay, Myanmar (H.H. Mon); Ministry Myanmar H9N2 viruses clustered with avian and human of Agriculture, Livestock and Irrigation, Nay Pyi Taw (K.N. Oo) H9N2 virus recovered in China in 2015. The phylogenetic analyses of internal protein genes showed similar find- DOI: https://dx.doi.org/10.3201/eid2306.161902 ings; the Myanmar H9N2 viruses were also closely related to avian H9N2 viruses from China. Five internal protein Routine surveillance of influenza A virus was conducted in genes, polymerase basic protein 2 (PB2), polymerase basic Myanmar during 2014–2015. Influenza A(H9N2) virus was protein 1 (PB1), polymerase acidic protein (PA), nucleo- isolated in Shan State, upper Myanmar. Whole-genome protein (NP), and membrane protein (M), were grouped sequencing showed that H9N2 virus from Myanmar was closely related to H9N2 virus of clade 4.2.5 from China. with the viruses of G1 lineage, and the nonstructural (NS)

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1041 RESEARCH LETTERS

Figure. Phylogenetic tree of H9 gene of influenza A(H9N2) viruses from Myanmar and reference viruses. Phylogenetic trees were constructed by using MEGA version 6.0 (http://www.megasoftware.net/) and a neighbor-joining algorithm with the Kimura 2-parameter model and 1,000 replications of bootstrap analysis. Only bootstrap numbers >70% are shown. Black circles represent isolates from this study, and open circles represent human H9N2 isolates. Virus clades are indicated at right. Scale bar indicates nucleotide substitutions per site.

1042 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS gene was grouped in BJ-94–like lineage (online Technical support to the Center of Excellence for Emerging and Re-emerg- Appendix Figure 3). In addition, BLAST analysis (https:// ing Infectious Diseases in Animals and the graduate scholarship blast.ncbi.nlm.nih.gov/Blast.cgi) of the 8 gene segments of for Neighboring Countries to T.N.L. The Thailand Research the Myanmar H9N2 viruses showed high percentages of Fund (TRF) provided financial support to the TRF Senior nucleotide identities with H9N2 viruses from China (online Scholar (RTA5780006) to A.A. Technical Appendix Table 2). Dr. Lin, a veterinarian, is a PhD candidate at the Center of Genetic analysis showed that the Myanmar H9N2 vi- Excellence for Emerging and Re-emerging Diseases in Animals, ruses possessed R-S-S-R at the HA cleavage site, indicating Faculty of Veterinary Science, Chulalongkorn University, low pathogenic characteristics (6). At the HA receptor bind- Bangkok, Thailand. His research interests are molecular ing sites, all Myanmar H9N2 isolates carried Q226L substi- epidemiology of emerging infectious diseases, especially tution, showing the affinity to humanα -2,6-glycan receptors influenza, in animals. (7). The Myanmar H9N2 viruses possessed all 7 HA receptor binding sites, identical to human H9N2 virus A/Zhong- shan/201501/2015 (online Technical Appendix Table 3) (8). References For NA gene analysis, all Myanmar H9N2 viruses had a 3-aa 1. Butt KM, Smith GJ, Chen H, Zhang LJ, Leung YH, Xu KM, et al. Human infection with an avian H9N2 influenza A virus in deletion (positions 62 to 64) in the NA stalk region, suggest- Hong Kong in 2003. J Clin Microbiol. 2005;43:5760–7. ing virus adaptation from wild birds to poultry (8). http://dx.doi.org/10.1128/JCM.43.11.5760-5767.2005 Internal protein gene analysis showed that the Myan- 2. Shanmuganatham K, Feeroz M, Jones-Engel L, Smith G, mar H9N2 isolates possessed both avian- and human-spe- Fourment M, Walker D, et al. Antigenic and molecular characterization of avian influenza A(H9N2) viruses, cific amino acids. For example, no mutation was observed Bangladesh. Emerging Infect Dis. 2013;19:1393–1402 . at E627 and D701 in PB2, retaining the avian character- http://dx.doi.org/10.3201/eid1909.130336 istics (9), whereas human-specific amino acids were ob- 3. Gao R, Cao B, Hu Y, Feng Z, Wang D, Hu W, et al. Human served at 13P in PB1 and 409N in PA (10). Amino acids infection with a novel avian-origin influenza A (H7N9) virus. N Engl J Med. 2013;368:1888–97. http://dx.doi.org/10.1056/ concerning virulence of the H9N2 viruses were also indi- NEJMoa1304459 cated. For example, the PA gene carried 672L and the NS1 4. Zhang Y, Yin Y, Bi Y, Wang S, Xu S, Wang J, et al. Molecular gene contained 149A, relating to increased virulence of the and antigenic characterization of H9N2 avian influenza virus virus (10). All Myanmar H9N2 viruses also possessed 31N isolates from chicken flocks between 1998 and 2007 in China. Vet Microbiol. 2012;156:285–93. http://dx.doi.org/10.1016/ in the M2 gene, suggesting amantadine resistance. Our re- j.vetmic.2011.11.014 sults, based on the whole-genome analysis, showed that the 5. Liu S, Ji K, Chen J, Tai D, Jiang W, Hou G, et al. Panorama Myanmar H9N2 viruses possessed all genetic signatures phylogenetic diversity and distribution of Type A influenza virus. and virulence determinants similar to avian and human PLoS One. 2009;4:e5022. http://dx.doi.org/10.1371/journal. pone.0005022 H9N2 viruses circulating in China in 2015 (online Techni- 6. Ge F-F, Zhou J-P, Liu J, Wang J, Zhang W-Y, Sheng L-P, et al. cal Appendix Tables 3–5). Genetic evolution of H9 subtype influenza viruses from live poultry In conclusion, the Myanmar H9N2 viruses were con- markets in Shanghai, China. J Clin Microbiol. 2009;47:3294–300. sidered of low pathogenicity in chickens. However, public Pub Med http://dx.doi.org/10.1128/JCM.00355-09 7. Matrosovich MN, Krauss S, Webster RG. H9N2 influenza A viruses health concerns should be raised because the viruses are from poultry in Asia have human virus-like receptor specificity. closely related tom and possess virulence determinants Virology. 2001;281:156–62. http://dx.doi.org/10.1006/viro.2000.0799 similar to, human H9N2 reported in China in 2015. The 8. Zhao Y, Li S, Zhou Y, Song W, Tang Y, Pang Q, et al. human cases likely represent a spillover of the virus to hu- Phylogenetic analysis of hemagglutinin genes of H9N2 avian influenza viruses isolated from chickens in Shandong, China, mans from poultry (8). The Myanmar H9N2 viruses were between 1998 and 2013. BioMed Res Int. 2015;2015:267520. obtained from healthy chickens in LBMs near the China– http://dx.doi.org/ 10.1155/2015/267520 Myanmar border, where sources of poultry are mostly from 9. Yuan J, Xu L, Bao L, Yao Y, Deng W, Li F, et al. Characterization China. Because LBMs are the major sources of several sub- of an H9N2 avian influenza virus from aFringilla montifringilla brambling in northern China. Virology. 2015;476:289–97. types of influenza viruses, this environment provides op- http://dx.doi.org/10.1016/j.virol.2014.12.021 portunities for influenza viruses to mix, be transmitted, and 10. Kandeil A, El-Shesheny R, Maatouq AM, Moatasim Y, exchange their gene segments. Infection and transmission Shehata MM, Bagato O, et al. Genetic and antigenic evolution in humans and poultry can frequently occur in LBM set- of H9N2 avian influenza viruses circulating in Egypt between 2011 and 2013. Arch Virol. 2014;159:2861–76. http://dx.doi.org/ tings. Therefore, public awareness, control measures, and 10.1007/s00705-014-2118-z routine disease surveillance should be implemented. Address for correspondence: Alongkorn Amonsin, Department of This research was supported by the National Research Universi- Veterinary Public Health, Faculty of Veterinary Science, ty Fund (NRU59-028-HR) and the research chair grant NSTDA Chulalongkorn University, Bangkok 10330, Thailand; email: (P-15-50004). Chulalongkorn University provided financial [email protected]

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1043 RESEARCH LETTERS

PCR Detection of Mimivirus available in the public domain since March 2016, the date on which the article by Zhang et al. was submitted (3). Our in silico analysis found that none of the 25 genomes avail- Didier Raoult, Anthony Levasseur, able from lineage A and 7 genomes from lineage B were Bernard La Scola complementary with their primers and probes, and only 5 Author affiliation: Aix-Marseille Université, Marseille, France of 19 lineage C genomes showed sufficient sequence similarity to permit successful amplification (Table). Overall, DOI: https://dx.doi.org/10.3201/eid2306.161896 the PCR used by these authors would be predicted to detect only ≈26% of lineage C mimiviruses and <10% of pub- In an epidemiological study, mimivirus was reported as an lished mimivirus genomes overall. unlikely cause of human respiratory infections in China. Our As proposed by Ngounga et al. (4), the great variabil- analysis revealed the nonsensitivity of the PCR method, ity across the 3 lineages of mimivirus genomes requires which detected less than 10% of the current known mimivi- the design of lineage-specific primers to improve PCR rus. We conclude that epidemiologic studies must use ac- sensitivity (4). Until PCRs are improved, mimiviruses curate and sensitive laboratory test methods. might be best detected by metagenomic approaches that have recently seen success in terms of mimivirus detec- he article by Zhang et al. (1) reporting the failure to tion. For example, mimivirus-like reads have been identi- Tdetect mimivirus (Acanthamoeba polyphaga mimi- fied by metagenomics in 1 (8%) of 12 fecal samples from virus) in patients with acute respiratory symptoms was children with acute diarrhea (5); 5 (27%) of 18 prostatic flawed in 2 ways. First, a single mimivirus was detected secretion samples from persons with prostatitis (6); all by PCR but was dismissed by the authors as colonization. DNA libraries tested from the plasma of patients with Second, and of greater concern, the PCR used by the au- hepatitis (7); 0.0018% of total viral reads from nasopha- thors would not detect most known mimiviruses because ryngeal samples from 210 patients with respiratory tract of primer/probe mismatches. In fact, PCRs used in most symptoms (8); and all tested fecal, mid-vaginal, buccal mimivirus prevalence studies would fail to detect most mucosal, and retroauricular crease samples, with a pre- mimiviruses, given the extensive genome sequence diver- dominance of mimiviruses in fecal samples reaching sity of these viruses (2). ≈33% of the total viral reads (9). We evaluated the PCR primers/probes used by Zhang In metagenomics approaches, the number of reads de- et al. (1) for sequence similarity with 51 mimivirus genomes tected and the distribution along the whole viral genome

Table. Mimivirus genomes in the public domain screened to determine whether the genomes were complementary with their primers/probes* Lineage A (GenBank accession no.) Lineage B (GenBank accession no.) Lineage C (GenBank accession no.) Mimivirus fauteuil (CWKE00000000) Moumouvirus ochan (CXPE00000000) Megavirus shan (CWIL00000000) Mimivirus longchamps (CXOV00000000) Moumouvirus istres (PRJEB9189) Megavirus montpellier3 (CWIA00000000) Mimivirus marais (PRJEB9199) Moumouvirus battle49 (CXOS00000000) Megavirus mont1 (CWIO00000000) Mimivirus lactours (CXOL00000000) Moumouvirus saoudian (CXOQ00000000) Megavirus avenue9 (CWHN00000000) Mimivirus PointeRouge1 (CXOW00000000) Moumouvirus boug1 (PRJEB9190) Megavirus balcon (PRJEB9194) Mimivirus PointeRouge2 (CXOU00000000) Moumouvirus goulette (KC008572) Megavirus ursino (CWJY00000000) Mimivirus T3 (CXOR00000000) Moumouvirus Monve (JN885994– Megavirus T1 (CWIB00000000) Mimivirus huitre A06 (CWJT00000000) JN886001) Megavirus battle43 (CWHV00000000) Mimivirus amazonia (CWHO00000000) Megavirus T4 (CWJR00000000) Mimivirus battle7 (CWJW00000000) Megavirus J3 (CWIN00000000) Mimivirus battle19 (CWJX00000000) Megavirus T6 (CWJQ00000000) Mimivirus battle27 (CWJS00000000) Megavirus chilensis (NC_016072) Mimivirus battle57 (CWJZ00000000) Megavirus courdo11 (JX975216) Mimivirus battle66 (CWKA00000000) Megavirus lba† (JX885207) Mimivirus battle83 (CWKG00000000) Powai lake megavirus (KU877344) Mimivirus battle86 (CWKD00000000) Megavirus terra1 (KF527229) Mimivirus T2 (CXOT00000000) Megavirus courdo5 (CWID00000000) Mimivirus battle6 (CWJU00000000) Megavirus bus (CWIQ00000000) Mimivirus dakar4 (CWKF00000000) Megavirus courdo7 (JN885990– Mimivirus (AY653733) JN885993) Mimivirus Bombay (KU761889) Mimivirus terra 2 (PRJNA240235) Samba virus (KF959826) Mimivirus lentille (AFYC00000000) Mimivirus urine† *Bold indicates genomes that would be predicted to be successfully amplified using the PCR primers and probes in the article by Zhang et al. (1). †Mimiviruses isolated from human samples.

1044 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017

RESEARCH LETTERS are essential parameters for viral detection. In practice, the Autochthonous Case of search for mimivirus is complicated by the great genetic variability of the virus and the restricted availability of Eosinophilic Meningitis mimivirus culture systems to a few research laboratories Caused by Angiostrongylus (10). The deficiencies we found in the report by Zhang et cantonensis, France, 2016 al. highlight the need for carefully designed epidemiologic studies using sensitive laboratory test methods to accurately assess mimivirus prevalence and the potential role of Yann Nguyen, Benjamin Rossi, Nicolas Argy, mimivirus in human disease. Catherine Baker, Beatrice Nickel, Hanspeter Marti, Virginie Zarrouk, Dr. Raoult is an Associate Editor of Emerging Infectious Sandrine Houzé, Bruno Fantin, Agnès Lefort Diseases and professor of microbiology at the School of Medicine, Aix-Marseille University, Marseille, France, and is Author affiliations: Hôpital Beaujon, Clichy, France (Y. Nguyen, head of the University Hospital Institute Méditerranée Infection B. Rossi, C. Baker, V. Zarrouk, B. Fantin, A. Lefort); Hôpital in Marseille. Bichat, Paris, France (N. Argy, S. Houzé); Swiss Tropical and Public Health Institute, Basel, Switzerland (B. Nickel, H. Marti); References University of Basel, Basel (B. Nickel, H. Marti) 1. Zhang XA, Zhu T, Zhang PH, Li H, Li Y, Liu EM, et al. Lack of mimivirus detection in patients with respiratory disease, China. DOI: https://dx.doi.org/10.3201/eid2306.161999 Emerg Infect Dis. 2016;22:2011–2. http://dx.doi.org/10.3201/ eid2211.160687 2. Colson P, Aherfi S, La Scola B, Raoult D. The role of giant viruses We report a case of a 54-year-old Moroccan woman living of amoebas in humans. Curr Opin Microbiol. 2016;31:199–208. in France diagnosed with eosinophilic meningitis caused http://dx.doi.org/10.1016/j.mib.2016.04.012 by Angiostrongylus cantonensis. Diagnosis was based on 3. Levasseur A, Bekliz M, Chabrière E, Pontarotti P, La Scola B, clinical symptoms and confirmed by testing of serum and Raoult D. MIMIVIRE is a defence system in mimivirus that cerebrospinal fluid samples. Physicians should consider the confers resistance to virophage. Nature. 2016;531:249–52. risk for A. cantonensis infection outside of endemic areas. http://dx.doi.org/10.1038/nature17146 4. Ngounga T, Pagnier I, Reteno DG, Raoult D, La Scola B, Colson P. Real-time PCR systems targeting giant viruses of ngiostrongylus cantonensis is a rat lungworm that has long amoebae and their virophages. Intervirology. 2013;56:413–23. http://dx.doi.org/10.1159/000354563 Abeen recognized as a cause of eosinophilic meningitis in 5. Finkbeiner SR, Allred AF, Tarr PI, Klein EJ, Kirkwood CD, Southeast Asia, the Pacific Islands, and the Caribbean, where Wang D. Metagenomic analysis of human diarrhea: viral it is endemic (1). Although sporadic imported cases have been detection and discovery. PLoS Pathog. 2008;4:e1000011. described in European travelers (2–5), no autochthonous case http://dx.doi.org/10.1371/journal.ppat.1000011 6. Smelov V, Bzhalava D, Arroyo Mühr LS, Eklund C, of eosinophilic meningitis caused by A. cantonensis worms Komyakov B, Gorelov A, et al. Detection of DNA viruses in has been reported previously in metropolitan France. prostate cancer. Sci Rep. 2016;6:25235. http://dx.doi.org/10.1038/ A 54-year-old Moroccan woman was admitted to an srep25235 emergency ward in Paris in 2016 because of fever and head- 7. Law J, Jovel J, Patterson J, Ford G, O’Keefe S, Wang W, et al. Identification of hepatotropic viruses from plasma using deep ache lasting 2 weeks. She had a history of type 2 diabetes treat- sequencing: a next generation diagnostic tool. PLoS One. ed with metformin and did not receive any other medication 2013;8:e60595. http://dx.doi.org/10.1371/journal.pone.0060595 before the onset of symptoms. She was a pesco-vegetarian and 8. Lysholm F, Wetterbom A, Lindau C, Darban H, Bjerkner A, worked as a cleaning woman in an office in Paris. She had Fahlander K, et al. Characterization of the viral microbiome in patients with severe lower respiratory tract infections, using taken up residence near Paris in the 1980s and had not trav- metagenomic sequencing. PLoS One. 2012;7:e30875. eled out of France since, except for Morocco over 2 years ago. http://dx.doi.org/10.1371/journal.pone.0030875 At admission, her blood pressure was 126/68 mm Hg, 9. Rampelli S, Soverini M, Turroni S, Quercia S, Biagi E, Brigidi P, and her pulse was regular at 79 beats/min. The physical et al. ViromeScan: a new tool for metagenomic viral community profiling. BMC Genomics. 2016;17:165. http://dx.doi.org/10.1186/ examination revealed meningeal signs with neck stiffness s12864-016-2446-3 and photophobia without neurologic localization signs. No 10. Saadi H, Pagnier I, Colson P, Cherif JK, Beji M, other clinical abnormality could be evidenced. Laboratory Boughalmi M, et al. First isolation of mimivirus in a patient testing of serum showed a leukocyte count of 12.1 × 109 with pneumonia. Clin Infect Dis. 2013;57:e127–34. 9 http://dx.doi.org/10.1093/cid/cit354 cells/L (reference range 4.5–11.0 × 10 /L) with 18% (2.2 × 109 cells/L) eosinophils and a C-reactive protein of 73 Address for correspondence: Didier Raoult, Faculté de Médecine, mg/dL (reference range 0.08–3.1 mg/L). Cerebrospinal Université de la Méditerranée, URMITE UMR CNRS 6236 IRD 198, fluid (CSF) analysis showed a leukocyte count of 950 cells/ Centre National de Référence, 27 Blvd Jean Moulin, Marseille 13005 µL with 56% eosinophils, a glucose concentration within CEDEX 05, France; email: [email protected] reference range at 0.4 g/L, and an elevated protein level

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1045 RESEARCH LETTERS

of 0.9 g/L. CSF pressure was 16 cm H2O (0.16 kPa). CSF in France by consuming imported contaminated food, but PCR tests for viral diseases (including Epstein-Barr virus she did not declare any consumption of imported snails or and herpes simplex viruses 1 and 2); Gram stains and cul- freshwater shrimp. She only declared eating vegetables and tures for bacteria, mycobacteria, and fungi; and a search for fish bought in a local supermarket. We failed to identify the malignant cells were all negative. Serologic tests specific source of infection or risky behaviors. Because of global for HIV and human T-lymphotrophic viruses 1 and 2 were food trade, physicians should be aware of the risk for A. negative. Parasitologic examinations of stool samples did cantonensis infection even in patients who do not report not reveal any ova, cysts, or Charcot-Leyden crystals. Se- recent travel to endemic areas. rologic tests of peripheral blood samples for Toxocara spp., Trichinella spp., Schistosoma spp., Taenia solium cysticer- Acknowledgments cosis, Gnathostoma spp., and Fasciola hepatica, as well We thank Michèle Felce for assisting with the study and Karin as serologic tests of CSF samples for Gnathostoma, were Stoll-Rudin for expert technical assistance. all negative. In contrast, antibodies directed toward the A. Dr. Nguyen is a resident in internal medicine at the Assistance cantonensis 31-kDa antigen were detected by Western blot Publique–Hôpitaux de Paris. His primary research interests are both in serum and CSF samples. Computed tomography immunology and infectious diseases. (CT) and magnetic resonance imaging of the brain showed normal findings. Thoracic and abdominal CT showed a liver hypodensity without signs of malignancy. The patient References was treated by evacuation of the CSF, prednisone (1 mg/kg 1. Wang Q-P, Lai D-H, Zhu X-Q, Chen X-G, Lun Z-R. Human angiostrongyliasis. Lancet Infect Dis. 2008;8:621–30. equivalents for 1 wk), and albendazole (800 mg/d for 5 d), http://dx.doi.org/10.1016/S1473-3099(08)70229-9 which lead to the complete regression of symptoms. 2. Ali AB, Van den Enden E, Van Gompel A, Van Esbroeck M. Eosinophilic meningitis is defined as a CSF eosinophil Eosinophilic meningitis due to Angiostrongylus cantonensis in a count above 10% of the total cell count, or exceeding 10 eo- Belgian traveller. Travel Med Infect Dis. 2008;6:41–4. http://dx.doi.org/10.1016/j.tmaid.2007.09.001 sinophils/µL. The main etiologies of eosinophilic meningitis 3. Luessi F, Sollors J, Torzewski M, Müller HD, Siegel E, Blum J, are parasitic infectious diseases (among which A. cantonen- et al. Eosinophilic meningitis due to Angiostrongylus cantonensis sis and Gnathostoma spp. infections are the most frequent), in Germany. J Travel Med. 2009;16:292–4. http://dx.doi.org/ hematologic or neoplastic disorders, adverse drug reactions, 10.1111/j.1708-8305.2009.00337.x 4. Bärtschi E, Bordmann G, Blum J, Rothen M. Eosinophilic meningitis and primary eosinophilic meningitis (6). In this case, diag- due to Angiostrongylus cantonensis in Switzerland. Infection. nosis of angiostrongyliasis was confirmed by Western blot. 2004;32:116–8. http://dx.doi.org/10.1007/s15010-004-3028-x According to the literature, detection of the 31-kDa band by 5. Malvy D, Ezzedine K, Receveur M-C, Pistone T, Crevon L, immunoblot shows a high sensitivity and specificity (>99% Lemardeley P, et al. Cluster of eosinophilic meningitis attributable to Angiostrongylus cantonensis infection in French policemen for both) for the diagnosis of A. cantonensis infections (7). troop returning from the Pacific Islands. Travel Med Infect Dis. Human infection by A. cantonensis lungworm results 2008;6:301–4. http://dx.doi.org/10.1016/j.tmaid.2008.06.003 from the ingestion of uncooked paratenic hosts (freshwa- 6. Lo Re V III, Gluckman SJ. Eosinophilic meningitis. Am J Med. ter shrimps, crabs, and frogs); intermediate hosts (snails 2003;114:217–23. http://dx.doi.org/10.1016/S0002-9343(02)01495-X 7. Eamsobhana P, Gan XX, Ma A, Wang Y, Wanachiwanawin D, and slugs); or poorly cleaned contaminated vegetables that Yong HS. Dot immunogold filtration assay (DIGFA) for the rapid have been in contact with these hosts (8). After passage detection of specific antibodies against the rat lungworm through the gastrointestinal tract, the larvae cross into sys- Angiostrongylus cantonensis (Nematoda: Metastrongyloidea) temic circulation and migrate to the central nervous system. using purified 31-kDa antigen. J Helminthol. 2014;88:396–401. http://dx.doi.org/10.1017/S0022149X13000321 The incubation period of this disease varies from 7 to 35 8. Ramirez-Avila L, Slome S, Schuster FL, Gavali S, Schantz PM, days (1). Although angiostrongyliasis is mainly a tropical Sejvar J, et al. Eosinophilic meningitis due to Angiostrongylus disease, this infection has become an emerging infection and Gnathostoma species. Clin Infect Dis. 2009;48:322–7. in Southeast Asia, the Pacific Islands, and the Caribbean http://dx.doi.org/10.1086/595852 9. Cowie RH, Hollyer JR, da Silva AJ, Hollingsworth RG, Dixon MC, Islands and in travelers who return home from endemic ar- Eamsobhana P, et al. Workshop on research priorities for eas (9). Some authors have suggested that global warming, management and treatment of angiostrongyliasis. Emerg Infect Dis. changes in dietary habits, increased global transportation of 2012;18:e1. http://dx.doi.org/10.3201/eid1812.120499 food products, increased international trade, and traveling 10. Eamsobhana P. Eosinophilic meningitis caused by Angiostrongylus cantonensis—a neglected disease with escalating importance. Trop could explain this expansion (10). Biomed. 2014;31:569–78. Because the patient we describe had not left the Paris area for 2 years, an infection by consumption of contaminated Address for correspondence: Yann Nguyen, Department of Internal food in Morocco can be excluded; A. cantonensis larvae Medicine, Hôpital Beaujon, Assistance Publique – Hôpitaux de Paris, do not survive longer than a few weeks to a few months in 100 boulevard du Général Leclerc, 92110 Clichy, France; email: the human body (8). She might have acquired the infection [email protected]

1046 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS

Zika Virus–Associated throat, headache, rash and joint pain resolved; however, 8 days after symptom onset, the neuropsychiatric symptoms Cognitive Impairment in worsened and included excessive energy, decreased sleep, Adolescent, 2016 rapid and tangential speech, grandiose thinking, impulsiv- ity, decreased inhibition, and behavioral regression suggestive of hypomania. The patient’s psychiatrist prescribed an Jason Zucker, Natalie Neu, Claudia A. Chiriboga, antipsychotic medication. An ambulatory and 24-hour elec- Veronica J. Hinton, Marc Leonardo, Arif Sheikh, troencephalogram and magnetic resonance imaging of the Kiran Thakur brain without contrast showed no abnormalities. Author affiliations: Columbia University Medical Center, New York, Symptoms failed to improve, and 13 days after symp- New York, USA (J. Zucker, N. Neu, C.A. Chiriboga, V.J. Hinton, tom onset, the patient was hospitalized for further work-up. A. Sheikh, K. Thakur); Child, Adolescent, and Adult Psychiatry Fourteen days after symptom onset, magnetic resonance private practice, New York (M. Leonardo) imaging of the brain with contrast showed no abnormalities. Serum was positive for dengue virus IgG and IgM (on- DOI: https://dx.doi.org/10.3201/eid2306.162029 line Technical Appendix Table 2). Repeat Zika virus testing was not performed because of the prior positivity of urine Incidence of neurologic manifestations associated with Zika RT-PCR. Fifteen days after symptom onset, cerebrospinal virus infection has been increasing. In 2016, neuropsycho- fluid (CSF) obtained by lumbar puncture contained 8 leu- logical and cognitive changes developed in an adolescent kocytes/μL, 1,000 erythrocytes/μL, 55 mg/dL glucose, and after travel to a Zika virus–endemic area. Single-photon 30 mg/dL protein. CSF testing for Zika virus, performed emission computed tomography and neuropsychological at the Wadsworth Laboratory, New York City Department testing raised the possibility that Zika virus infection may of Mental Health and Hygiene (New York, NY, USA), was lead to neuropsychiatric and cognitive symptoms. positive for IgM but negative by RT-PCR. After discharge, the patient’s symptoms of regres- lthough the full clinical spectrum of complications as- sion, disinhibition, and cognitive impairment persisted. Asociated with Zika virus disease remains unclear, evi- Seven weeks after symptom onset, single-photon emis- dence of an association between Zika virus infection and sion computed tomography revealed mildly heteroge- diseases of the nervous system is growing (1). Increased neous cerebral cortical perfusion, with focal moderate- incidence of neurologic manifestations has been reported, to-severe hypoperfusion in the inferior left frontal region including cerebral birth abnormalities (2), Guillain-Barré (online Technical Appendix Figure 2); neuropsychologi- Syndrome (3), meningoencephalitis (4), and memory loss cal testing demonstrated evidence of superior intellectual (5). We describe a case of probable central nervous system ability (Wechsler Intelligence Scale for Children–fourth (CNS) infection with Zika virus associated with the onset of edition, General Ability Index), probably reflecting func- neuropsychological and cognitive changes in an adolescent. tion before illness (online Technical Appendix Table 3). In summer 2016, an adolescent traveled to a Zika vi- Processing speed was significantly slowed relative to oth- rus–endemic island in the southern Caribbean, where many er skills; the Processing Speed Index score was >2 SDs insect bites occurred. Before the trip, the patient’s medical lower than the General Ability Index score. Performance history included mild depression treated with a selective on most memory tests and tests of executive function serotonin reuptake inhibitor. Assessment after return to was within normal limits yet lower than expected given the United States indicated that mental and physical health reported performance before illness. Immediate and de- were the same as before travel. One week after return, the layed visual recall of the Rey complex figure was poor, patient experienced sore throat, headache, diffuse scarlati- reflecting primary difficulty encoding new visual- infor niform rash (online Technical Appendix Figure 1, https:// mation. On a standardized behavioral questionnaire, the wwwnc.cdc.gov/EID/article/23/6/16-2029-Techapp1.pdf), patient self-reported psychiatric symptoms including anx- joint pain, confusion, and short-term memory loss; no fever iety, racing thoughts, and an inability to turn off thoughts. or eye inflammation were noted. The patient had not experienced racing thoughts before Three days after symptom onset, the patient received the Zika virus infection, and the anxiety symptoms had care from a pediatrician; extensive laboratory testing re- worsened since the infection. In addition, the patient revealed urine reverse transcription PCR (RT-PCR) results ported significant and functionally limiting fatigue. Nine positive for Zika virus. Serologic results were positive weeks after onset, symptoms were better but not resolved; for Epstein-Barr virus, consistent with the patient’s his- because of concerns that these symptoms were triggered tory of infection 15 months earlier (online Technical Ap- by a postinfectious immune-mediated process, a trial of pendix Table 1). Five days after symptom onset, the sore intravenous immunoglobulin was administered. Fifteen

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1047 RESEARCH LETTERS weeks after symptom onset, the patient’s symptoms were Highly Pathogenic Avian better but not fully resolved. Although we cannot entirely rule out Epstein-Barr vi- Influenza Virus (H5N8) rus as a possible trigger, the length of time between previous Clade 2.3.4.4 Infection in infection and onset of neuropsychological symptoms would Migratory Birds, Egypt be unusual. In addition, although it is impossible to exclude contributions of coinfection from other mosquitoborne viruses (e.g., dengue and chikungunya), given the Zika virus Abdullah A. Selim, Ahmed M. Erfan, positivity on RT-PCR, the patient’s condition met criteria Naglaa Hagag, Ali Zanaty, Abdel-Hafez Samir, for definitive Zika virus infection and the CSF IgM titer was Mohamed Samy, Ahmed Abdelhalim, consistent with CNS involvement of Zika virus. The changes Abdel-Satar A. Arafa, Mohamed A. Soliman, on single-photon emission computed tomographs and neuro- Momtaz Shaheen, Essam M. Ibraheem, psychological test scores raise the possibility that Zika virus Ibrahim Mahrous, Mohamed K. Hassan, infection may trigger neuropsychiatric and cognitive symp- Mahmoud M. Naguib toms. Although we cannot prove that the patient’s symptoms Author affiliations: National Laboratory for Veterinary Quality were related to Zika virus, clinicians should be aware of this Control on Poultry Production, Animal Health Research Institute, potential association and the value of closely monitoring pa- Giza, Egypt (A.A. Selim, A.M. Erfan, N. Hagag, A. Zanaty, tients with Zika virus infection. A.-H. Samir, M. Samy, A. Abdelhalim, A.-S.A. Arafa, M.A. Soliman, K.T., who helped edit the manuscript, is a World Health M. Shaheen, E.M. Ibraheem, M.K. Hassan, M.M. Naguib); Organization consultant on neurologic manifestations in the General Organization for Veterinary Services, Giza (I. Mahrous); context of the Zika virus outbreak, receives an honorarium as Friedrich-Loeffler-Institut, Federal Research Institute for Animal chief consult editor on Zika virus for Medscape consults, Health, Insel-Riems, Germany (M.M. Naguib) and participates in Zika research through the Neuroviruses DOI: https://dx.doi.org/10.3201/eid2306.162056 Emerging in the Americas Study. K.T. receives funding support from the National Institutes of Health, and J.Z. receives funding We isolated highly pathogenic avian influenza virus (H5N8) support from a National Institutes of Health training grant of clade 2.3.4.4 from the common coot (Fulica atra) in (T32 AI007531). Egypt, documenting its introduction into Africa through migratory birds. This virus has a close genetic relationship Dr. Zucker is a postdoctoral fellow in adult and pediatric with subtype H5N8 viruses circulating in Europe. Enhanced infectious diseases at Columbia University Medical Center. His surveillance to detect newly emerging viruses is warranted. research interests include improving prevention and treatment for adolescents and young adults living with, or at risk for, HIV infection, hepatitis C, and sexually transmitted diseases. vian influenza is a highly contagious disease of poul- Atry that continues to spread across the globe in bird populations. Occasionally, transmission of a highly patho- References genic avian influenza virus (HPAIV) from infected poultry 1. Araujo AQ, Silva MT, Araujo AP. Zika virus-associated neurological disorders: a review. Brain. 2016;139:2122–30. to humans results in a severe public health crisis (1). http://dx.doi.org/10.1093/brain/aww158 In 2010, strains of HPAIV (H5N8) of clade 2.3.4.4 2. Mlakar J, Korva M, Tul N, Popović M, Poljšak-Prijatelj M, were first detected among wild birds in Asia and later Mraz J, et al. Zika virus associated with microcephaly. N Engl J spread to domestic birds across China, South Korea, and Med. 2016;374:951–8. http://dx.doi.org/10.1056/NEJMoa1600651 3. dos Santos T, Rodriguez A, Almiron M, Sanhueza A, Ramon P, Japan (2,3). Most recently, a novel reassortant virus of sub- de Oliveira WK, et al. Zika virus and the Guillain-Barré Syndrome— type H5N8 clade 2.3.4.4 was reported in Russia and further case series from seven countries. N Engl J Med. 2016;375:1598– spread to many countries in Europe, Asia, and the Middle 601. http://dx.doi.org/10.1056/NEJMc1609015 East (4,5). The spread of HPAIV (H5N8) strains has been 4. Carteaux G, Maquart M, Bedet A, Contou D, Brugières P, Fourati S, et al. Zika virus associated with meningoencephalitis. N Engl J linked to the overlapping flyways of migratory wild birds Med. 2016:374:1595–6. http://dx.doi.org/10.1056/NEJMc1602964 that come from different continents; this mingling of wild 5. Nicastri E, Castilletti C, Balestra P, Galgani S, Ippolito G. Zika birds poses a major concern worldwide (4,6). virus infection in the central nervous system and female genital Egypt is one of the most notable migration spots for tract. Emerg Infect Dis. 2016;22:2228–30. http://dx.doi.org/ 10.3201/eid2212.161280 migratory birds crossing Europe, Asia, and Africa. In early winter each year, thousands of migrating waterfowl use Address for correspondence: Jason Zucker, Columbia University Egypt as a resting stop before they continue their journey Medical Center, 622 West 168th St, PH4 West–474, New York, NY southward through the African continent through the East 10032, USA; email: [email protected] Africa/East Asia and Mediterranean/Black Sea migratory

1048 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS

Figure. Structural and phylogenetic modeling of highly pathogenic avian influenza virus (H5N8), EG-CA285, from migratory birds, Egypt, 2016. A) Three-dimensional structural homology model for the hemagglutinin protein of EG-CA285 created by using the ancestral virus of clade 2.3.4.4b (A/duck/Zhejiang/6D18/2013 [H5N8]) as a template. Amino acids distinguishing the EG-CA285 sequence from the modeling template are shown in red; green depicts unique mutations distinguishing this virus from the virus detected in summer 2016 in Russia, A/great crested grebe/Uvs-Nuur-Lake/341/2016. B) Phylogenetic tree of the nucleotide sequences of avian influenza virus hemagglutinin genes. Maximum-likelihood calculations were done with IQ-TREE software (http://iqtree.cibiv.univie.ac.at/) under the best- fit model according to the Akaike criterion (general time reversible plus gamma plus G4 model). Bold indicates strains from Egypt; gray shading indicates strains currently circulating in Europe. Scale bar indicates nucleotide substitutions per site.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1049 RESEARCH LETTERS flyways. Lake Manzala in northern Egypt is a source of features of the HA and NA gene segments confirm the in- fish and a major refuge for many migratory birds (online tercontinental dissemination of HPAIV (H5N8) through Technical Appendix Figure 1, http://wwwnc.cdc.gov/EID/ wild birds and its introduction into Egypt. article/23/6/16-2056-Techapp1.pdf1). During the evolution of subtype H5Nx viruses of During a targeted surveillance for avian influenza clade 2.3.4.4, frequent reassortment has been noted with viruses (AIVs) conducted in migratory birds by Com- other co-circulating HPAIVs and low pathogenicity AIVs munity Animal Health Outreach (CAHO) program on in different countries in Europe, North America, and East November 24, 2016, we collected 19 oropharyngeal and Asia (8). Strains of HPAIV (H5N8) have been involved in cloacal swab samples from diseased (mild depression) and multiple independent reassortment events with other AIV dead migratory birds (common coot, Fulica atra; pintail subtypes found in wild birds in China, South Korea, the ducks, Anas acuta; and Garganey ducks, A. querquedula) United States, and recently in Russia (5,9). The probable in a live bird and fish market in the Damietta Governor- introduction of HPAIV (H5N8) to poultry populations in ate in Egypt. (Hunted migratory birds are commonly sold Egypt will further complicate disease control and preven- for food in markets in this region.) Two samples from 2 tion, especially if HPAIV (H5N1) of clade 2.2.1.2 and low common coots were confirmed positive for AIV and were pathogenicity AIV (H9N2) strains of G1 lineage are enzo- subtyped as H5N8 by using specific real-time reverse tran- otic in poultry (10). scription quantitative PCR (RT-qPCR) (online Technical In addition, the threat of emergence of a novel re- Appendix). On November 30, 2016, the identification of assortants with unpredictable gene constellations of HPAIV (H5N8) from 2 common coots was reported to the HPAIV (H5N8) strains with enzootic strains of AIV is World Organisation for Animal Health. Notably, this newly a public health concern. Therefore, we recommend en- emerged HPAIV (H5N8) was detected in Egypt in the same hanced surveillance to quickly detect newly emerged place, Damietta Governorate, where HPAIV (H5N1) was viruses. Commercial and backyard poultry owners must first identified in 2006 during the global spread of HPAIV follow the recommended biosecurity measures. The de- (H5N1) viruses of clade 2.2 (7). Immediately thereafter, ac- tection and immediate reporting of novel HPAIV (H5N8) tive targeted surveillance for AIV was conducted around strains in Egypt will help increase AIV surveillance, de- Lake Manzala and the surroundings areas for AIV that in- tection, and prevention preparedness in other countries cluded wild birds and domestic poultry; however, no more of continental Africa. positive cases were detected. We successfully isolated and characterized 1 HPAIV Acknowledgments (H5N8) strain by nucleotide sequencing and phylogenetic We acknowledge Timm Harder and Anandan Paldurai for their analyses on the basis of its hemagglutinin (HA) and neur- fruitful discussions and helpful comments on this written work. aminidase (NA) gene segments. The isolate was named A/ We thank the colleagues in the National Laboratory for common coot/Egypt/CA285/2016 (EG-CA285). Veterinary Quality Control on Poultry Production, General The amino acid sequence of the protease cleavage Organizations for Veterinary Services, and Community site of EG-CA285 HA protein revealed multiple basic Animal Healt=h Outreach team for their valuable support. amino acids, PLREKRRKR/GLF, which is characteristic We are grateful to those who submitted sequence data to the of HPAIV. The receptor-binding pocket of EG-CA285 HA GISAID database (online Technical Appendix Table 1). protein showed markers of avian receptor–specific binding: Q222 and G224. We observed 3 amino acid assignment Dr. Selim is the technical manager of the National Laboratory for differences in the HA protein, namely, R22K, E268G, and Quality Control on Poultry Production, Animal Health Research D487Y, which distinguished EG-CA285 from the recent Institute, Egypt. His research interest includes diagnosis and HPAIV (H5N8) clade 2.3.4.4b strain isolated in Russia (A/ molecular epidemiology of avian influenza viruses. great-crested-grebe/Uvs-Nuur-Lake/341/2016; GISAID accession no EPI_ISL_224580) (Figure, panel A). In the References NA protein, we observed 4 substitution mutations (V8A, 1. Alexander DJ, Brown IH. History of highly pathogenic avian V31L, G126E, I407T) that distinguished the EG-CA285 influenza. Rev Sci Tech. 2009;28:19–38. http://dx.doi.org/ from the subtype found in Russia. Phylogenetic analysis of 10.20506/rst.28.1.1856. 2. Wu H, Peng X, Xu L, Jin C, Cheng L, Lu X, et al. Novel HA and NA gene sequences revealed that EG-CA285 virus reassortant influenza A(H5N8) viruses in domestic ducks, eastern is clustered with clade 2.3.4.4b, along with the recent vi- China. Emerg Infect Dis. 2014;20:1315. http://dx.doi.org/10.3201/ ruses widely distributed throughout Europe (Figure, panel eid2008.140339. B; online Technical Appendix Figure 2). Even though the 3. Lee Y-J, Kang H-M, Lee E-K, Song B-M, Jeong J, Kwon Y-K, et al. Novel reassortant influenza A(H5N8) viruses, South Korea, unavailability of a full-length genomic sequence of this vi- 2014. Emerg Infect Dis. 2014;20:1087–9 and http://dx.doi.org/ rus is a limitation in this study, the genetic and phylogenetic 10.3201/eid2006.140233.

1050 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 RESEARCH LETTERS

4. European Centre for Disease Prevention and Control. Outbreak novel reassortants, United States, 2014–2015. Emerg Infect Dis. of highly pathogenic avian influenza A(H5N8) in Europe—18 2016;22:1283–5. http://dx.doi.org/10.3201/eid2207.160048. November 2016. Stockholm: The Centre; 2016. 9. Dalby A. Complete analysis of the H5 hemagglutinin and N8 5. Lee DH, Sharshov K, Swayne DE, Kurskaya O, Sobolev I, neuraminidase phylogenetic trees reveals that the H5N8 subtype Kabilov M, et al. Novel reassortant clade 2.3.4.4 avian has been produced by multiple reassortment events [version 1; influenza A(H5N8) virus in wild aquatic birds, Russia, 2016. referees: 1 approved with reservations]. [cited 2017 Apr 4]. Emerg Infect Dis. 2017;23:359–60. http://dx.doi.org/10.3201/ https://f1000research.com/articles/5-2463/v1 eid2302.161252. 10. Naguib MM, Arafa AS, El-Kady MF, Selim AA, Gunalan V, 6. Lee DH, Torchetti MK, Winker K, Ip HS, Song CS, Swayne DE. Maurer-Stroh S, et al. Evolutionary trajectories and diagnostic Intercontinental spread of Asian-origin H5N8 to North America challenges of potentially zoonotic avian influenza viruses H5N1 through Beringia by migratory birds. J Virol. 2015;89:6521–4. and H9N2 co-circulating in Egypt. J Mol Epidemiol Evol Gen http://dx.doi.org/10.1128/JVI.00728-15. Infect Dis. 2015;34:278–91. http://dx.doi.org/10.1016/ 7. Saad MD, Ahmed LS, Gamal-Eldein MA, Fouda MK, Khalil F, j.meegid.2015.06.004. Yingst SL, et al. Possible avian influenza (H5N1) from migratory bird, Egypt. Emerg Infect Dis. 2007;13:1120–1 and http://dx.doi.org/ Address for correspondence: Mahmoud M. Naguib, Friedrich-Loeffler- 10.3201/eid1307.061222. Institut Bundesforschungsinstitut fur Tiergesundheit Südufer 10, 8. Lee DH, Bahl J, Torchetti MK, Killian ML, Ip HS, DeLiberto TJ, et al. Highly pathogenic avian influenza viruses and generation of Greifswald–Insel Riems 17493, Germany; email: [email protected] October 2016: Disease Patterns • Community- and Healthcare- • Distinct Zika Virus Lineage in Associated Clostridium difficile Salvador, Bahia, Brazil Infections, Finland, 2008–2013 • Streptococcus suis Serotype 2 • Carbapenem Resistance in Capsule In Vivo Clonally Distinct Clinical Strains • Estimation of Severe MERS-CoV of Vibrio fluvialis Isolated from Cases in the Middle East, Diarrheal Samples 2012–2016 • Whole-Genome Characterization • Hypervirulent Clone of Group B of Epidemic Neisseria meningitidis Streptococcus Serotype III Sequence Serogroup C and Resurgence of Type 283, Hong Kong, 1993–2012 Serogroup W in Niger, 2015 • Outbreaks of Human Salmonella Infections Associated with Live Poultry, USA, 1990–2014 • Vaccine-Derived Polioviruses and Children with Primary Immunodeficiency, Iran, 1995–2014 • Infection-Related Deaths from Refractory Juvenile Idiopathic Arthritis • Ebola Virus Disease in Children, • Accuracy of Diagnosis of Sierra Leone, 2014–2015 Human Granulocytic Anaplasmosis in China • Systematic Review and Meta-Analysis • Population-Level Effects of Human of the Treatment Efficacy • Chikungunya Virus in Febrile Papillomavirus Vaccination of Doxycycline for Rectal Humans and Aedes aegypti Programs on Infection with Lymphogranuloma Venereum in Mosquitoes, Yucatan, Mexico Nonvaccine Human Men who have Sex with Men • Daily Reportable Disease Papillomavirus Genotypes • Increase in Meningococcal Spatiotemporal Cluster • Cat-Scratch Disease in the United Serogroup W Disease, Victoria, Detection, New York, New York, States, 2005–2013 Australia, 2013–2015 USA, 2014–2015 https://wwwnc.cdc.gov/eid/articles/ issue/22/10/table-of-contents

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1051 ABOUT THE COVER

Ramón y Cajal (1852–1934), Calyx of Held (3676) (detail), circa 1920. Pencil, India ink, ink washes on cardboard or paper. Digital image courtesy of the Cajal Institute-CSIC, Madrid, Spain.

“Unexplored Continents and Great Stretches of Unknown Territory”

Byron Breedlove

his month’s cover art could be, at first glance, an im- Santiago Ramón y Cajal, one of the most influential neuro- Tage conjured by the vision of a surrealist painter in the scientists of all time. 1920s or an intricate detail lifted from a textile designed Cajal was born in Petilla de Aragón, a small village by British artist William Morris in the 1870s. It borrows in northeastern Spain. He is well known for many key facets from both but is neither. Rather, this striking image contributions to our knowledge of the structure of the is one among the thousands of scientific images created by nervous system and especially of the brain and spinal cord. Cajal’s career as a researcher, teacher, and mentor Author affiliation: Centers for Disease Control and Prevention, spanned more than 50 years, and he published numer- Atlanta, Georgia, USA ous groundbreaking books and more than 100 articles DOI: https://dx.doi.org/10.3201/eid2306.AC2306 in French and Spanish scientific periodicals. He was

1052 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 ABOUT THE COVER also an artist, photographer, bodybuilder, chess player, Cajal’s detailed work studying and documenting the and publisher. complexities of the central nervous system and its myriad A eureka moment in Cajal’s career occurred in Ma- pathways, connections, and function remains essential to drid in 1887 when another scientist showed him the Golgi current research focused on understanding and treating in- stain that made microscopic structures in brain tissue fections of the central nervous system. This system may be samples appear black against an amber background. Al- either the main target or a secondary target for infections though Golgi’s process was imprecise and inconsistent, that can result in result in lethal, catastrophic, and debilitat- Cajal realized that this method could be his pathway to ing illness. studying microscopic details of the nervous system. Af- The catalog of bacteria, viruses, fungi, parasites, and ter he refined and stabilized the black reaction, he logged prions responsible for central nervous system disorders long hours at his microscope painstakingly scrutinizing spans the alphabet from Angiostrongylus to Zika virus. central nervous system tissues from animals and humans. Though we have learned a great deal about how to detect Cajal eventually made more than 3,000 exquisitely de- and treat central nervous system infections, the global bur- tailed sketches of these images, and scientists and educa- den of death and illness underscores the value of public tors are still using them. health efforts to prevent, diagnose, and treat them. As Cajal EID's cover art is Cajal’s drawing of the calyces of said, “The brain is a world consisting of a number of unex- Held—so named because Hans Held described the struc- plored continents and great stretches of unknown territory.” ture’s resemblance to floral calyces in his 1893 article— which is one of the largest synapses in the human brain Acknowledgment stem and is integral to the ability to detect and localize We gratefully recognize the Instituto Cajal (CSIC) which high-frequency sounds. Cajal portrayed the calyces as in- generously permitted the reproduction of the Cajal original dividual cells, encircled by thick black lines that show the drawing that forms part of the Cajal Legacy, ©CSIC. axons. From this perspective, the calyces resemble flowers or seedpods intertwined about a central stalk. Each dot, dash, or line, and the precise patterns, tones, Bibliography 1. Andres-Barquin PJ. Santiago Ramón y Cajal and the Spanish and density in the image are in keeping with Cajal’s credo school of neurology. Lancet Neurol. 2002;1:445–52. that “A graphic representation of the object observed guar- http://dx.doi.org/10.1016/S1474-4422(02)00192-8 antees the exactness of the observation itself.” (Note that 2. Garcia-Lopez P, Garcia-Marin V, Freire M. The histological slides the handwritten “Manzano number” in the upper left en- and drawings of Cajal. Front Neuroanat. 2010;4:9. 3. Gibson WC. Ramón y Cajal and his school: personal recollections. circled by the stamped words “Museo Cajal Madrid” is a J Hist Med Allied Sci. 1994;49:546–64. http://dx.doi.org/10.1093/ cataloguing convention added by the librarian who inven- jhmas/49.4.546 toried Cajal’s archive in the 1940s.) 4. Hamilton J. Art exhibition celebrates drawings by the founder of In The Beautiful Brain: The Drawings of Santiago modern neuroscience [cited 2017 Apr 12] http://www.npr.org/ sections/health-shots/2017/01/26/511455876/art-exhibition- Ramón y Cajal, Lyndel King and Eric Himmel describe celebrates-drawings-by-the-founder-of-modern-neuroscience how Cajal developed his sketches: “He preferred to work 5. Held H. The central auditory system [in German]. Arch Anat freehand, rarely resorting to a camera lucida, a device that Physiol Anat Abt. 1893;17:201–48. projects the image from a microscope onto paper where it 6. John CC, Carabin H, Montano SM, Bangirana P, Zunt JR, Peterson PK. Global research priorities for infections that can be traced. He might start a drawing in pencil, and then affect the nervous system. Nature. 2015;527:S178–86. later go over it in India ink, adding ink washes or watercol- http://dx.doi.org/10.1038/nature16033 ors for tonal areas.” 7. Nobel Prize. Santiago Ramón y Cajal – Biographical [cited Cajal’s many distinctions include winning, along with 2017 Apr 12] http://www.nobelprize.org/nobel_prizes/medicine/ laureates/1906/cajal-bio.html the Italian histologist Camillo Golgi, the Nobel Prize in 8. Parikh V, Tucci V, Galwankar S. Infections of the nervous Physiology or Medicine in 1906. His legacy endures through system. Int J Crit Illn Inj Sci. 2012;2:82–97. http://dx.doi.org/ the continued work and influence of the Spanish school 10.4103/2229-5151.97273 of neurology in Madrid, which developed from his work 9. Swanson L, King L, Himmel E, editors. The beautiful brain: the drawings of Santiago Ramón y Cajal. New York: Abrams; 2017. and vision. King and Himmel offered this assessment of p 11–18, 21–28, 34, 84, 104–105, 193. Cajal’s scientific sketches: “When we look at his drawings today, we see not diagrams or arguments, but the first clear Address for correspondence: Byron Breedlove, EID Journal, Centers pictures of that remote frontier, drawn by the man who for Disease Control and Prevention, 1600 Clifton Rd NE, Mailstop C19, traveled farthest into its endless reaches.” Atlanta, GA 30329-4027, USA; email: [email protected]

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1053 NEWS AND NOTES

Upcoming Infectious Disease Activities June 1–5, 2017 ASM American Society for Microbiology New Orleans, LA, USA Upcoming Issue http://www.showsbee.com/fairs/ • Risk Factors for Legionella longbeachae Legionnaires’ Disease, 25161-ASM-Microbe-2017.html New Zealand June 4–8, 2017 Council of State and Territorial • Novel Retinal Lesion in Ebola Survivors, Sierra Leone, 2016 Epidemiologists • Case−Control Study of Risk Factors for Meningococcal Disease in Chile 2017 Annual Conference Boise, ID, USA • Phylogeography of Burkholderia pseudomallei Isolates, http://www.csteconference.org/2017/ Western Hemisphere June 19–21, 2017 • Effects of Zika Virus Strain and Aedes Mosquito Species on Transmission of Respiratory Viruses Vector Competence Harbour Grand Hong Kong • Clonal Clusters and Virulence Factors of Group C and G Streptococci Hong Kong, China https://transmission2017.med.hku.hk/ Causing Severe Infections, Manitoba, Canada, 2012–2014 mass_email.html • Attributable Fraction of Influenza Virus Detection to Illness in June 20–21, 2017 HIV-Infected and HIV-Uninfected Patients, South Africa, 2012–2016 SHEA/CDC ORTP Regional • Competence of Aedes aegypti, Ae. albopictus, and Training Program Culex quinquefasciatus Mosquitoes as Zika Virus Vectors, China Sonesta Philadelphia Rittenhouse Square Philadelphia, PA, USA • One-Year Follow-Up Antibody Responses in Patients Infected with http://ortp.shea-online.org/ MERS-CoV, South Korea, 2015 December 5–8, 2017 • Concurrent Infection with Hepatitis C Virus and Streptococcus 6th National Congress of Tropical pneumoniae, Alberta, Canada Medicine and International Symposium on HIV/aids Infection • Emergency Meningococcal ACWY Vaccination Program for Teenagers 9th National Congress of Microbiology to Control Group W Meningococcal Disease, England, 2015–2016 and Parasitology • Francisella tularensis ssp. holarcticain Australian Ringtail Possums 80th Anniversary of the Institute of Tropical Medicine Pedro Kourí • Environmental Factors as Key Determinants for Visceral Havana, Cuba Leishmaniasis in Solid Organ Transplant Recipients, Madrid, Spain http://microbiologia2017.sld.cu/ index.php/microbiologia/2017 • Rabbit Hepatitis E Virus Infections in Humans, France February 1–3, 2018 • Detection and Genetic Characterization of Human Adenovirus 14 8th Advances in Aspergillosis Strain in Students with Influenza-Like Illness, New York, 2014–2015 Lisbon, Portugal • Norovirus GII.P16/GII.2–Associated Gastroenteritis, China, 2016 www.AAA2018.org • Association of GII.P16-GII.2 Recombinant Norovirus Strain with March 1–4, 2018 18th International Congress Increased Norovirus Outbreaks, Guangdong, China, 2016 on Infectious Diseases (ICID) • Escherichia coli Isolates with mcr-1 Gene, Australia, 2011 and 2013 Buenos Aires, Argentina http://www.isid.org/icid/ • Live Cell Therapy as Potential Risk Factor for Q Fever Announcements • Mycobacterium gordonae in Patient with Facial Ulcer, Nosebleeds, To submit an announcement, send an email message to EIDEditor ([email protected]). Include and Positive T-SPOT.TB Test, China the date of the event, the location, the sponsoring organization(s), and a website that readers may visit • Rickettsia sibirica mongolitimonae Infection, Turkey, 2016 or a telephone number or email address that readers may contact for more information. Complete list of articles in the July issue at Announcements may be posted on the journal Web http://www.cdc.gov/eid/upcoming.htm page only, depending on the event date.

1054 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Earning CME Credit To obtain credit, you should first read the journal article. After reading the article, you should be able to answer the following, related, multiple-choice questions. To complete the questions (with a minimum 75% passing score) and earn continuing medical education (CME) credit, please go to http://www.medscape.org/journal/eid. Credit cannot be obtained for tests completed on paper, although you may use the worksheet below to keep a record of your answers. You must be a registered user on http://www.medscape.org. If you are not registered on http://www.medscape.org, please click on the “Register” link on the right hand side of the website. Only one answer is correct for each question. Once you successfully answer all post-test questions, you will be able to view and/or print your certificate. For questions regarding this activity, contact the accredited provider, CME@medscape. net. For technical assistance, contact [email protected]. American Medical Association’s Physician’s Recognition Award (AMA PRA) credits are accepted in the US as evidence of participation in CME activities. For further information on this award, please go to https://www.ama-assn.org. The AMA has determined that physicians not licensed in the US who participate in this CME activity are eligible for AMA PRA Category 1 Credits™. Through agreements that the AMA has made with agencies in some countries, AMA PRA credit may be acceptable as evidence of participation in CME activities. If you are not licensed in the US, please complete the questions online, print the AMA PRA CME credit certificate, and present it to your national medical association for review. Article Title Sporadic Creutzfeldt-Jakob Disease in 2 Plasma Product Recipients, United Kingdom

CME Questions 1. Your patient is a 62-year-old woman with a long A. Cerebrospinal fluid real-time quaking-induced history of clotting disorder and recent onset of conversion (CSF RT-QuIC) was positive in only rapidly progressive dementia with prominent motor 1 patient features. According to the case reports by Urwin and B. Neuropathology was typical for sCJD in both patients, colleagues, which of the following statements about and neither had evidence of peripheral pathogenesis the clinical features of 2 cases of sporadic Creutzfeldt- on immunostaining of lymphoreticular tissues as is Jakob disease (sCJD) reported in patients with clotting seen in variant CJD (vCJD) disorders treated with fractionated plasma products is C. Both patients had an MM genotype at codon 129 of most accurate? PRNP, proving the diagnosis of sCJD A. Course was rapidly progressive with symptoms D. Only 1 patient had a type 1A isoform PrPSc on including incoordination, shuffling gait, seizure, Western blot daytime hypersomnolence, myoclonus, emotional 3. According to the case reports by Urwin and lability, dysphasia, and limb rigidity colleagues, which of the following statements about B. Vision was unaffected the clinical implications of 2 cases of sCJD reported C. One patient had a history of potential iatrogenic in patients with clotting disorders treated with exposure, and the other had a family history of CJD fractionated plasma products is correct? D. Survival duration from time of onset was roughly 1 year A. The study proves that treatment with plasma products causes sCJD 2. According to the case reports by Urwin and B. Look-back studies in the United States and United colleagues, which of the following statements Kingdom have shown transfusion-transmission about the laboratory and pathology findings of of sCJD 2 cases of sCJD reported in patients with clotting C. Plasma-derived products pose no risk for sCJD disorders treated with fractionated plasma products transmission is correct? D. It is essential to continue to search for cases of transfusion-transmission of sCJD through CJD surveillance programs

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 1055 Earning CME Credit To obtain credit, you should first read the journal article. After reading the article, you should be able to answer the following, related, multiple-choice questions. To complete the questions (with a minimum 75% passing score) and earn continuing medical education (CME) credit, please go to http://www.medscape.org/journal/eid. Credit cannot be obtained for tests completed on paper, although you may use the worksheet below to keep a record of your answers. You must be a registered user on http://www.medscape.org. If you are not registered on http://www.medscape.org, please click on the “Register” link on the right hand side of the website. Only one answer is correct for each question. Once you successfully answer all post-test questions, you will be able to view and/or print your certificate. For questions regarding this activity, contact the accredited provider, CME@medscape. net. For technical assistance, contact [email protected]. American Medical Association’s Physician’s Recognition Award (AMA PRA) credits are accepted in the US as evidence of participation in CME activities. For further information on this award, please go to https://www.ama-assn.org. The AMA has determined that physicians not licensed in the US who participate in this CME activity are eligible for AMA PRA Category 1 Credits™. Through agreements that the AMA has made with agencies in some countries, AMA PRA credit may be acceptable as evidence of participation in CME activities. If you are not licensed in the US, please complete the questions online, print the AMA PRA CME credit certificate, and present it to your national medical association for review. Article Title Relative Risk for Ehrlichiosis and Lyme Disease in an Area Where Vectors for Both Are Sympatric, New Jersey, USA

CME Questions 1. You are seeing a 23-year-old previously healthy C. A. americanum has historically been found in the woman for a 2-day history of fever and headache. She southeastern United States spent last week hiking in the mountains, and you are D. Samples of A. americanum have been declining in concerned that she may have a tick-borne illness. New Jersey Which of the following statements regarding the clinical presentation and prognosis of ehrlichiosis is 3. Ehrlichiosis is in your differential diagnosis for this most accurate? patient. What was the predicted ratio of ehrlichiosis to Lyme disease in the current study? A. It frequently presents with more general symptoms such as fever and headache A. 0.05 B. Rash is necessary to make the clinical diagnosis of B. 0.6 ehrlichiosis C. 1.0 C. Ehrlichiosis is not associated with any organ damage D. 2.3 D. Mortality rates associated with ehrlichiosis may be as high as 20% 4. How did the actual rate of reported ehrlichiosis compare with that in the statistical model in the 2. You obtain a more thorough history of tick current study? exposures from this patient. What should you A. The actual rate of ehrlichiosis was approximately consider regarding the vectors for ehrlichiosis and 100% lower than that in the model Lyme disease? B. The actual rate of ehrlichiosis was approximately 20% A. Ixodes scapularis can transmit both Borrelia lower than that in the model burgdorferi and Ehrlichia chaffeensis C. The actual rate of ehrlichiosis was approximately the B. I. scapularis is much more aggressive than same as that in the model Amblyomma americanum D. The actual rate of ehrlichiosis was approximately 50% higher than that in the model

1056 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 23, No. 6, June 2017 Emerging Infectious Diseases is a peer-reviewed journal established expressly to promote the recognition of new and reemerging infectious diseases around the world and improve the understanding of factors involved in disease emergence, prevention, and elimination. The journal is intended for professionals in infectious diseases and related sciences. We welcome contributions from infectious disease specialists in academia, industry, clinical practice, and public health, as well as from specialists in economics, social sciences, and other disciplines. Manuscripts in all categories should explain the contents in public health terms. For information on manuscript categories and suitability of proposed articles, see below and visit http://wwwnc.cdc.gov/eid/pages/author-resource-center.htm.

Summary of Authors’ Instructions biographical sketch of first author or of both authors if only 2 authors. This section comprises case series papers and concise reviews of infectious diseases or closely Author’s Instructions. For a complete list of EID’s manuscript guidelines, see the related topics. Preference is given to reviews of new and emerging diseases; how- author resource page: http://wwwnc.cdc.gov/eid/page/author-resource-center. ever, timely updates of other diseases or topics are also welcome. If detailed methods are included, a separate section on experimental procedures should immediately fol- Manuscript Submission. To submit a manuscript, access Manuscript Central from low the body of the text. the Emerging Infectious Diseases web page (www.cdc.gov/eid). Include a cover letter indicating the proposed category of the article (e.g., Research, Dispatch), verifying the Research. Articles should not exceed 3,500 words and 50 references. Use of sub- word and reference counts, and confirming that the final manuscript has been seen and headings in the main body of the text is recommended. Photographs and illustrations are approved by all authors. Complete provided Authors Checklist. encouraged. Provide a short abstract (150 words), 1-sentence summary, and biographical sketch. Report laboratory and epidemiologic results within a public health perspective. Manuscript Preparation. For word processing, use MS Word. Set the document Explain the value of the research in public health terms and place the findings in a larger to show continuous line numbers. List the following information in this order: title page, perspective (i.e., “Here is what we found, and here is what the findings mean”). article summary line, keywords, abstract, text, acknowledgments, biographical sketch, references, tables, and figure legends. Appendix materials and figures should bein Policy and Historical Reviews. Articles should not exceed 3,500 words and 50 refer- separate files. ences. Use of subheadings in the main body of the text is recommended. Photographs and illustrations are encouraged. Provide a short abstract (150 words), 1-sentence sum- Title Page. Give complete information about each author (i.e., full name, graduate mary, and biographical sketch. Articles in this section include public health policy or his- degree(s), affiliation, and the name of the institution in which the work was done). Clearly torical reports that are based on research and analysis of emerging disease issues. identify the corresponding author and provide that author’s mailing address (include phone number, fax number, and email address). Include separate word counts for abstract and text. Dispatches. Articles should be no more than 1,200 words and need not be divided into sections. If subheadings are used, they should be general, e.g., “The Study” and Keywords. Use terms as listed in the National Library of Medicine Medical “Conclusions.” Provide a brief abstract (50 words); references (not to exceed 15); figures Subject Headings index (www.ncbi.nlm.nih.gov/mesh). or illustrations (not to exceed 2); tables (not to exceed 2); and biographical sketch. Dis- Text. Double-space everything, including the title page, abstract, references, tables, patches are updates on infectious disease trends and research that include descriptions and figure legends. Indent paragraphs; leave no extra space between paragraphs. After of new methods for detecting, characterizing, or subtyping new or reemerging pathogens. a period, leave only one space before beginning the next sentence. Use 12-point Times Developments in antimicrobial drugs, vaccines, or infectious disease prevention or elimi- New Roman font and format with ragged right margins (left align). Italicize (rather than nation programs are appropriate. Case reports are also welcome. underline) scientific names when needed. Another Dimension. Thoughtful essays, short stories, or poems on philosophical Biographical Sketch. Include a short biographical sketch of the first author—both issues related to science, medical practice, and human health. Topics may include sci- authors if only two. Include affiliations and the author’s primary research interests. ence and the human condition, the unanticipated side of epidemic investigations, or how References. Follow Uniform Requirements (www.icmje.org/index.html). Do not people perceive and cope with infection and illness. This section is intended to evoke use endnotes for references. Place reference numbers in parentheses, not super- compassion for human suffering and to expand the science reader’s literary scope. scripts. Number citations in order of appearance (including in text, figures, and tables). Manuscripts are selected for publication as much for their content (the experiences they Cite personal communications, unpublished data, and manuscripts in preparation or describe) as for their literary merit. Include biographical sketch. submitted for publication in parentheses in text. Consult List of Journals Indexed in Research Letters Reporting Cases, Outbreaks, or Original Research. EID Index Medicus for accepted journal abbreviations; if a journal is not listed, spell out publishes letters that report cases, outbreaks, or original research as Research Letters. the journal title. List the first six authors followed by “et al.” Do not cite references in Authors should provide a short abstract (50-word maximum), references (not to exceed the abstract. 10), and a short biographical sketch. These letters should not exceed 800 words in the Tables. Provide tables within the manuscript file, not as separate files. Use the MS main body of the text and may include either 1 figure or 1 table. Do not divide Research Word table tool, no columns, tabs, spaces, or other programs. Footnote any use of bold- Letters into sections. face. Tables should be no wider than 17 cm. Condense or divide larger tables. Extensive Letters Commenting on Articles. Letters commenting on articles should contain a tables may be made available online only. maximum of 300 words and 5 references; they are more likely to be published if submitted Figures. Submit editable figures as separate files (e.g., Microsoft Excel, PowerPoint). within 4 weeks of the original article’s publication. Photographs should be submitted as high-resolution (600 dpi) .tif or .jpeg files. Do not em- Commentaries. Thoughtful discussions (500–1,000 words) of current topics. bed figures in the manuscript file. Use Arial 10 pt. or 12 pt. font for lettering so that figures, Commentaries may contain references (not to exceed 15) but no abstract, figures, or symbols, lettering, and numbering can remain legible when reduced to print size. Place fig- tables. Include biographical sketch. ure keys within the figure. Figure legends should be placed at the end of the manuscript file. Books, Other Media. Reviews (250–500 words) of new books or other media on Videos. Submit as AVI, MOV, MPG, MPEG, or WMV. Videos should not exceed 5 emerging disease issues are welcome. Title, author(s), publisher, number of pages, and minutes and should include an audio description and complete captioning. If audio is other pertinent details should be included. not available, provide a description of the action in the video as a separate Word file. Published or copyrighted material (e.g., music) is discouraged and must be accompanied Conference Summaries. Summaries of emerging infectious disease conference ac- by written release. If video is part of a manuscript, files must be uploaded with manu- tivities (500–1,000 words) are published online only. They should be submitted no later script submission. When uploading, choose “Video” file. Include a brief video legend in than 6 months after the conference and focus on content rather than process. Provide the manuscript file. illustrations, references, and links to full reports of conference activities. Types of Articles Online Reports. Reports on consensus group meetings, workshops, and other activities in which suggestions for diagnostic, treatment, or reporting methods related to Perspectives. Articles should not exceed 3,500 words and 50 references. Use of infectious disease topics are formulated may be published online only. These should not subheadings in the main body of the text is recommended. Photographs and illustra- exceed 3,500 words and should be authored by the group. We do not publish official tions are encouraged. Provide a short abstract (150 words), 1-sentence summary, and guidelines or policy recommendations. biographical sketch. Articles should provide insightful analysis and commentary about Photo Quiz. The photo quiz (1,200 words) highlights a person who made notable new and reemerging infectious diseases and related issues. Perspectives may address contributions to public health and medicine. Provide a photo of the subject, a brief clue factors known to influence the emergence of diseases, including microbial adaptation and to the person’s identity, and five possible answers, followed by an essay describing the change, human demographics and behavior, technology and industry, economic devel- person’s life and his or her significance to public health, science, and infectious disease. opment and land use, international travel and commerce, and the breakdown of public health measures. Etymologia. Etymologia (100 words, 5 references). We welcome thoroughly researched derivations of emerging disease terms. Historical and other context could be included. Synopses. Articles should not exceed 3,500 words in the main body of the text or include more than 50 references. Use of subheadings in the main body of the Announcements. We welcome brief announcements of timely events of interest to text is recommended. Photographs and illustrations are encouraged. Provide a short our readers. Announcements may be posted online only, depending on the event date. abstract (not to exceed 150 words), a 1-line summary of the conclusions, and a brief Email to [email protected].